Un sistema circulatorio cerrado y de alta presión precisa de un

UNIVERSIDAD DE MURCIA

FACULTAD DE MEDICINA

Identification of new elements involved in

antithrombin regulation.

Identificación de nuevos elementos implicados en la

regulación de antitrombina.

Da. Mª Eugenia de la Morena Barrio

2013

D. Javier Corral de la Calle y D. Vicente Vicente García, Profesores

Doctores de la Universidad de Murcia del Área de Hematología en el Departamento de

Medicina Interna de la Universidad de Mucia,

CERTIFICAN:

Que la tesis Doctoral titulada “Identificación de nuevos elementos implicados en la

regulación de antitrombina”, realizada por Dª. Mª Eugenia de la Morena Barrio

bajo su inmediata dirección y supervisión, y que se presenta para la obtención del grado

de Doctor Internacional por la Universidad de Murcia.

En Murcia, a 8 de Enero de 2013

Fdo: Dr. Javier Corral de la Calle Fdo: Dr. Vicente Vicente García

U N I V E R S I D A D D E M U R C I ADEPARTAMENTO DE MEDICINA INTERNAFacultad de Medicina

U N I V E R S I D A D D E M U R C I AU N I V E R S I D A D D E M U R C I ADEPARTAMENTO DE MEDICINA INTERNAFacultad de Medicina



Los trabajos de investigación llevados a cabo en la presente memoria de Tesis

Doctoral han sido financiados por los siguientes proyectos:

“Investigación genómica, proteómica y terapéutica del sistema hemostático

en la enfermedad tromboembólica venosa y arterial”, proyecto SAF 2006-

06212 financiado por el Ministerio de Educación y Ciencia.

“Estudio genético y estructural de la antitrombina. Desarrollo de nuevas

drogas anticoagulantes”, proyecto otorgado por la Fundación Mutua

Madrileña.

“Anomalías de glicosilación en antitrombina. Implicaciones patológicas y

terapéuticas”, proyecto PI12/00657 financiado por el Instituto de Salud

Carlos III.

El firmante de esta memoria ha disfrutado de una Beca Nacional Predoctoral

de Formación e Investigación en la Salud (FI09/00190) del Ministerio de Ciencia e

Innovación.

En Murcia, a 8 de Enero de 2013

Fdo: Mª Eugenia de la Morena Barrio





Agradecimientos

Sois muchos los que formáis parte de esta tesis y a los que os debo también

mucho, por eso estas líneas me resultan especialmente difíciles aunque por otra parte

son las que más ilusión me hacen. A todos os doy las gracias.

En primer lugar, me gustaría expresar mi más sincero agradecimiento al Dr.

Vicente Vicente García, co-director de esta tesis y jefe del grupo, por darme la

oportunidad de iniciarme en el mundo de la investigación abriéndome las puertas de su

grupo. Es mi deber agradecerle la confianza que ha depositado en mí y sus constantes

esfuerzos por ofrecerme una formación científica de calidad. Quiero agradecerle sus

consejos, preocupaciones, disponibilidad, buen hacer, su calidad profesional y

capacidad de trabajo y todo ello transmitiendo siempre serenidad, tacto, interés y

sentido del humor.

En segundo lugar, con profundo agradecimiento me dirijo al Dr. Javier Corral,

director y principal motor de este trabajo. Espero que acepte mi reconocimiento no solo

a la dirección de esta tesis, sino también a la orientación y formación científica que le

debo. Su capacidad de trabajo, exprimiendo cada segundo de cada minuto, su espíritu

crítico, su entusiasmo, inquietud y motivación junto con su gran bondad humana,

mostrando interés más allá de lo profesional, con sus constantes consejos y

preocupaciones, son sus pilares fundamentales, por lo que cualquiera se sentiría

afortunadísimo de tenerlo como director de tesis, jefe, colega o amigo. Solo espero

haber asimilado algo de todo lo que me ha transmitido.

A Toñi Miñano le debo tanto… Gracias por esos ratos de intenso trabajo en el

laboratorio donde he trabajado muy a gusto. Orden, trabajo, disciplina, organización,

constancia, generosidad y un gran corazón. Siempre con una sonrisa en los labios y con

palabras acertadas de ánimo. Le agradezco su ejemplo, su amistad y sus constantes

desvelos casi maternales.

A las Dras. Anabel Antón e Irene Martínez quisiera agradecerles de forma

especial lo mucho que he recibido de ellas. A Anabel le debo la formación genómica de

mis primeros años. Le agradezco su afectuoso acogimiento y cariño durante esos años,

además de sus grandes dotes como pedagoga. A Irene mi formación más proteómica.

Entre tantas cosas siempre me ha transmitido orden y rigurosidad en el trabajo y el no

conformarme con la mediocridad. Les agradezco a las dos sus siempre buenos consejos,

sugerencias y cariño.

Al Dr. Constantino Martínez me gustaría agradecerle toda la formación en

biología molecular que me ha aportado y que ha sido fundamental en el desarrollo de

algunos apartados de la tesis. Además, le agradezco su gran disponibilidad y su

constante alegría creando un ambiente en el grupo muy agradable.

Al Dr. Pepe Rivera su espíritu crítico, su interés, atención y disponibilidad.

Agradecerle los conocimientos sobre hemostasia primaria (la plaqueta) que siempre nos

transmite.

A la Dra. Rocío González-Conejero su orden y trabajo acompañando siempre una

sonrisa y una gran serenidad.

A los Dres. Paqui Ferrer, Marisa Lozano, Mª José Candela, Francisco Ayala y

Gloria Soler por sus siempre acertadas sugerencias y espíritu crítico. A Marisa

agradecerle especialmente su valiosa ayuda con el citómetro y la formación en

hematología que me ha aportado.

A los Dres. José Anotonio Guerrero, Leyre, Adriana, Pepe Navarro, Virginia y

Raúl les agradezco su ejemplo y trayectoria. Han ido abriéndome camino y me han

ayudado siempre que lo he necesitado. Han sido mis puntos de referencia que he

intentado seguir hasta estos días. A José G me gustaría agradecerle además su tiempo

dedicado a la revisión del inglés del manuscrito, sus correcciones, observaciones y

sugerencias.

A Isabel y Sonia les agradezco los muchos y muy buenos momentos que hemos

compartido como becarias “precarias”. El día a día ha sido mucho más llevadero a su

lado, compartiendo y aprendiendo juntas.

A las nuevas incorporaciones Ginés, Salam, Eva, Ana y Sergio por ir renovando y

rejuveneciendo al grupo aportando nuevas ideas, visiones y modos de hacer.

A todo el maravilloso equipo técnico. A Lorena por todo lo que me ayudó los

primeros años, su cariño y buenos momentos que hemos compartido. A Jose Padilla le

agradezco todo su gran trabajo de secuenciación, su orden y organización. A Nuria le

agradezco la formación en biología celular y molecular. Siempre ha sido una gran ayuda

con los cultivos, las purificaciones de material genético, etc. A Natalia todo lo que me

ha aportado y ayudado en el laboratorio. A “su Rodri” le agradezcosu ayuda con la

impresión de la tesis. Y a Tina y Noelia sus favores y atenciones con las técnicas

genómicas.

Me gustaría mostrar también mi gratitud al resto de miembros del servicio de

Hematología y Oncología del Hospital Morales Meseguer y del Hospital Reina Sofía,

por su gran preparación y masa crítica que generan en las sesiones.

También agradezco a los enfermeros del Centro Regional de Hemodonación su

disponibilidad en la extracción de muestras de sangre.

En definitiva, mi agradecimiento a todos y cada uno de los que formáis parte de

este grupo por todo lo que transmite: profesionalidad, calidad humana, espíritu crítico,

disciplina, orden, saber hacer, exigencia, flexibilidad, rigurosidad, convivencia y por

supuesto, su gran potencial culinario. Su excelencia profesional combinada con su

excelencia humana han diluido los momentos duros, difíciles o de intenso trabajo, en

momentos amenos donde he disfrutado muchísimo.

Además, es mi deber agradecer al personal técnico y sanitario de los hospitales y

grupos con los que hemos colaborado, su interés y esfuerzo en el envío de muestras

biológicas de pacientes. Concretamente dar las gracias al Dr. Francisco España

(Hospital La Fe, Valencia), a la Dra. Carmen de Cos (Hospital Universitario Puerta del

Mar, Cádiz), al Dr. Jose Manuel Soria (Institut d’Investigació Sant Pau, Barcelona), al

Dr. Søren (Department of Haematology, Aalborg Hospital, Dinamarca), a la Dra. Teresa

Sevivas (Centro Hospitalar de Coimbra, Portugal), al Dr. Jaeken (Universitair

Ziekenhuis Gasthuisberg, Leuven, Belgium.) y al Dr. Dirk Lefeber (Institute for genetic

and metabolic disease, Holanda), así como a todos los pacientes que han participado en

este estudio. A todos ellos mi más profundo respeto y reconocimiento.

A los profesores Jaak Jaeken, Kathleen Freson, Marc Hoylaerts y Michela de

Michele agradecerles además los meses de estancia en su departamento (Center for

Molecular and Vascular Biology, KU Leuven, Leuven, Bélgica) y la interesante

colaboración que hemos abierto desde entonces. Especialmente agradecer a la Dra.

Michela y al Dr. Marc su cariño y esfuerzos por hacer productiva mi estancia allí.

A Vanessa Ferreiro, presidenta de la Asociación de enfermos con trastornos de

glicosilación congénita, mostrarle mi admiración por la labor que desempeña y mi

gratitud por hacer posible la colaboración con el grupo del profesor Jaeken.

Al Dr. Francisco España y a la Dra. Silvia Navarro quisiera agradecerles también

los días que pasé en su departamento en el Hospital La Fe de Valencia. Les agradezco

su ayuda con la técnica de generación de trombina.

A los profesores Marc Hoylaerts y Søren Risom Kristensen me gustaría

expresarles mi reconocimiento por su tiempo dedicado en la valoración de la tesis para

aspirar a la mención de Doctorado Internacional.

Y lo más importante: mi familia y amigos. Creo que no hace falta que me extienda

mucho aquí. Bien saben ellos lo mucho que agradezco lo que han sido y son para mí,

estos años y siempre.

Por último, debo agradecer al Instituto de Salud Carlos III la beca que he

disfrutado estos cuatro años y que me ha hecho posible llegar hasta aquí.

A mi Abuelo (in memoriam)

A mis padres

Abbreviations

2D Two dimension electrophoresis

α1AT Alpha 1 antitrypsin

AB Aminobenzamide

ACN Acetonitrile

AMI Acute Myocardial Infarction

AT Antithrombin

bp Nucleotide base pairs

CDG Congenital Disorder of Glycosylation

CDT Carbohydrate-Deficient Transferrin

CS-DS Chondroitin-dermatan sulfate

kDa Kilo Dalton

DHB Dihydroxybenzoic acid

DVT Deep Venous Thrombosis

ELISA Enzyme Linked Immunosorbent Assay

FXa Activated Blood Coagulation Factor X

FPLC Fast Protein Liquid Chromatography

GAGs Glycosaminoglycans

GAIT Genetic Analysis of Idiopathic Thrombophilia

GPIbα Glycoprotein Ib alpha

GPIIIa Gycoprotein IIIa

GU Glucose Units

GWAS Genome Wide Association Study

HEK-EBNA Human Embryonic Kidney cells expressing the Epstein Barr

Nuclear Antigen 1

HepG2 Human hepatocellular liver carcinoma cell line

HILIC Hydrophilic Interaction Liquid Chromatography

HPLC High Performance Liquid Chromatography

HS Heparan Sulfate

INR International Normalized Ratio

LLO Lipid-linked oligosaccharide

MALDI-TOF Matrix Assited Laser Desorption Ionization-Time-of-flight

MPI Mannose-6-phosphate isomerase

MS Mass Spectrometry

PAI-1 Plasminogen activator inhibitor 1

PBS Phosphate Buffer Saline

PCR Polymerase chain reaction

pI Isoelectric point

PLC-PRF-5 Human hepatome cell line

PMM2 Phosphomannomutase 2

PVDF Polyvinylidene Fluoride

qRT-PCR quantitative Real Time PCR

Q-TOF Quadrupole-Time-of-Flight

ROS Reactive oxygen species

SDS-PAGE Sodium Dodecyl Sulfate Polyacrylamide Gel

SERPIN Serine protease inhibitor

siRNA small interfering RNA

SNP Single Nucleotide Polymorphism

TFA Trifluoroacetic acid

vWF von Willebrand Factor

General index

General index

General index

General index

Page

ABSTRACT 1

INTRODUCTION 5

1. Haemostatic system 7

2. Thrombophilic factors 8

3. Serpins 10

4. Antithrombin 13

4.1 Structure and biological function 14

4.2 Genetics 20

4.3 Interindividual variability of antithrombin 21

4.4 Antithrombin deficiency and venous thrombosis 21

OBJECTIVES 27

MATERIAL AND METHODS 31

1. Patients and healthy subjects 33

2. Blood sampling 34

3. Antithrombin levels and other thrombophilic tests 34

4. Amplification and sequencing of SERPINC1 34

5. Multiplex Ligation-dependent Probe Amplification (MLPA) 35

6. Genotyping of SERPINC1 genetic alterations 36

7. Prediction of transcription factor binding sites to the SERPINC1 promoter 37

8. Cloning and reporter assay 38

9. Genome wide association study 38

10. Validation study 39

11. Case-control study 40

12. Statistical analyses 40

13. LARGE and SERPINC1 gene expression 41

14. Purification of plasma antithrombin. Proteomic and glycomic analysis 41

15. LARGE gene silencing and effect on N-glycoproteins 43

16. Electrophoretic evaluation 44

17. N-glycosidase F, neuraminidase and galactosidase plasma treatment 44

General index

18. Purification and proteomic analysis of an antithrombin variant from a CDGIa-

like patient 45

19. Identification and quantification of transferrin glycoforms 45

20. PMM2 gene PCR amplification and sequencing 46

21. Prediction of splice-site, enhancers or silencers, and binding of splicing factors

in the PMM2 gene 47

22. Genotyping of IVS4 +21 G>C and IVS2 -13 G>A PMM2 mutations 47

23. Platelet purification and proteomic analysis 47

24. Treatment of HepG2 cells with proton and Ca2+ pump inhibitors and a pH

neutralizing agent 49

25. Epidemiologic studies of CDG-like disorders 49

RESULTS 51

OBJECTIVE 1

1. Anti-FXa activity variability in healthy population 53

2. Genetic variability of SERPINC1 in healthy subjects with extreme anti-FXa

activity, and validation of potential functional consequences 53

3. Genetic analysis in 25 patients with moderate antithrombin

deficiency 58

4. SERPINC1 analysis in subjects with classical antithrombin deficiency

carrying neither gross gene deletion nor mutation affecting exons or flanking

regions 59

OBJECTIVE 2

1. GWAS analysis. Association between genotype and plasma anti-FXa activity

in the GAIT study 61

2. Validation study 62

3. Thrombotic relevance of the LARGE rs762057 polymorphism 64

4. Functional studies 65

General index

OBJECTIVE 3

1. Diagnosis of a congenital disorder of glycosylation (PMM2-CDG) in a

patient with antithrombin deficiency and recurrent venous

thrombosis 71

2. Identification of a CDG-like disorder 74

3. Follow-up of patient P15 79

4. Proteomic analysis of platelet glycoproteins in PMM2-CDG and CDG-like

patients 79

5. Search for additional factors involved in CDG-like disorders 81

6. Epidemiologic analysis evaluating potential CDG-like disorders in patients

with thrombosis 82

DISCUSSION 85

CONCLUSIONS 105

BIBLIOGRAPHY 109

SPANISH ABSTRACT/RESUMEN EN CASTELLANO 125

APPENDIX I 133

APPENDIX II 149

General index

Index of tables and figures

Index of tables

and figures



Page

INTRODUCTION

Figure 1: Clot cascade, fibrinolysis and anticoagulant molecules. 8

Table 1: Frequent thrombophilic defects. 10

Figura 2: Native structure of α1-antitrypsin, the archetype structure of serpins (PDB: 1QLP s). 12

Figura 3: Inhibitory mechanism of Serpins. 13

Figura 4: Localization of disulfide bridges and N-glycosylation sites in antithrombin. 14

Figura 5: Composition of the N-glycans of antithrombin. 14

Figure 6: Main plasma glycoforms of antithrombin. The number and structural localization of N-glycans are shown. 15

Figure 7: Native and active forms of antithrombin. The interactions of the P1 residue (Arg393) in the native conformation are also shown. 16

Figure 8: Conformational changes on antithrombin induced by binding of heparin.17

Figure 9: Inhibition of FXa and thrombin by antithrombin. Relevance of the length of heparin. 18

Figure 10: Localization of the anticoagulant activity of antithrombin (AT) to the endothelial surface by heparan sulfates (HS). 19

Figure 11: Native and latent conformations of antithrombin (PDB: 1LT1FA and PDB: 2BEH, respectively). 20

Figure 12: SERPINC1 gen. Exons are marked as black boxes and Alu sequences with arrows. 21

Figure 13: Probability of thrombosis among carriers of different thrombophilic defects according to age. 22

MATERIAL AND METHODS

Table 2: Oligonucleotides used for PCR amplification of the SERPINC1 gene. 35

Figure 14: HRM-razor melting curves and electropherograms of wild type and heterozygous g.2085 T>C genotypes. 36

Table 3: Genotyping by PCR-allele specific restriction assay and High Resolution Melting. Mutated nucleotides are underlined. 37

Table 4: TaqMan® probes used for genotyping in the validation study. 39


Table 5: Oligonucleotides used for PCR amplification of the PMM2 gene. 46

RESULTS

OBJECTIVE 1

Figure 15: Distribution of anti-FXa activity in 307 Spanish blood donors. 53

Figure 16: Genetic variations identified in the SERPINC1 gene. 54

Table 6: Polymorphisms identified in healthy blood donors with low and high antithrombin levels. 55

Table 7: Plasma anti-FXa values of blood donors according to the genotype of SERPINC1 polymorphisms identified in selected 5 healthy subjects with extreme antithrombin levels. 56

Figure 17: Linkage disequilibrium (r2 values) between polymorphisms identified in our study. A gray scale is used to associate r2 values. 57

Figure 18: Haplotype analysis. 57

Table 8: Clinical and genetic features of patients with moderate antithrombin deficiency. 58

Figure 19: Identification and functional characterization of the g.2143 C>G modification. 60

OBJECTIVE 2

Figure 20: Manhattan plot GWAS with antithrombin phenotype. 61

Table 9: Single nucleotide polymorphisms (SNPs) associated with anti-FXa activity in the GWAS of the GAIT study and that were selected for validation studies. 62

Table 10: Genotype-phenotype analysis in the validation study. 63

Table 11: LARGE haplotypes identified in the validation study and their correlation

with anti-FXa activity. 63

Figure 21: Plasma anti-FXa activity according to the rs762057 genotype in the GAIT and validation studies. 64

Figure 22: Genotype and allelic frequencies of rs762057 in the case/control study including 849 patients with venous thrombosis and 862 controls. 65

Figure 23: Correlation between mononuclear LARGE expression and plasma anti-FXa activity (A) and antigen levels (B) in 10 healthy subjects. 66

Figure 24: Glycomic and proteomic analysis of healthy subjects with the highest (blue) and lowest (red) LARGE expression. 67


Figure 25: LARGE gene expression in HEK-EBNA and HepG2 cell lines 60 hours after transfection with LARGE siRNAs. 69

Figure 26: LARGE gene silencing in HepG2 and HEK-EBNA cell lines. 70

Figure 27: Identification of PMM2-CDG in the proband (P27). 73

Figure 28: Electrophoretic pattern of plasma antithrombin from a control and the proband (P27) compared to that of a mixture of alpha and beta antithrombin (AT) purified from plasma of healthy subjects. 73

Figure 29: Electropherograms of mutations identified in the PMM2 gene of the proband (P27). 74

Figure 30: Electrophoresis analysis in SDS-PAGE under reducing conditions of patients with a CDG-like disorder (P15, P16, P2 and P22). 75

Figure 31: HPLC profiles of plasma transferrin glycoforms of CDG-like patients (P15, P16 and P2), a control and a PMM2-CDG patient. 76

Figure 32: Q-TOF mass spectrometry of transferrin glycoprotein of CDG-like patients (P2, P15 and P16), a PMM2-CDG patient and a control. 78

Figure 33: Electropherograms of PMM2 mutations found in CDG-like patients (P15, P16 and P2). 78

Figure 34: Evolution of α1-antitrypsin glycoforms, plasma antithrombin (AT) activity (% of anti-FXa compared with a pool of 100 healthy controls) and asialo-transferrin (area of the asialotransferrin peak in HPLC) levels in patient P15 with CDG-like during five years of follow-up. 79

Figure 35: Western blot of platelet glycoproteins in a PMM2-CDG patient, a CDG-like patient (P15), and two controls. 80

Figure 36: 2D gel of platelet glycoproteins from a control, a PMM2-CDG patient, and a CDG-like patient (P15). 81

Figure 37: Electrophoretic analysis of antithrombin secreted to the conditioned medium of HepG2 treated with bafilomycin A1 (Baf), anmonium persulphate (NH4Cl) and cyclopiazonic acid (CPA). 82

Figure 38: Immunoflouresce of intracellular antithrombin from HepG2 cells treated anmonium persulphate (NH4Cl). 82

Table 12: Clinical and genetic features of patients and controls with a CDG-like pattern. 83

Figure 39: Prevalence of PMM2 p.Arg141His mutation in patients and controls. 84


DISCUSSION

Figure 40: Localization of g.2143 C>G in potential promoter regulatory regions of the SERPINC1 gene identified by DNase protection assay. 89

Table 13: Missense polymorphisms identified in the SERPINC1 gene. 91

Figure 41: N-glycosylation of proteins and its defects. 98

Figure 42: Potential interference of alcohol on the N-glycosylation metabolic

pathway. 102

Figure 43: Electrophoretic pattern of α1-Antitrypsin of an alcoholic patient comparing

with a control, PMM2-CDG and CDG-like patients. 103

Figure 44: HPLC profile of plasma transferrin glycoforms from an alcoholic

patient 103

Abstract

1

ABSTRACT

Abstract

2

Abstract

3

Antithrombin is a key anticoagulant serpin. Any factor affecting, even

moderately, the levels or function of this molecule will unbalance a finely regulated

haemostatic system, increasing the risk of thrombotic disorders, as widely demonstrated

with congenital antithrombin deficiency. Mutations in the coding regions of the gene

encoding antithrombin, SERPINC1, have been identified in a high proportion of patients

with antithrombin deficiency. However, there are few studies evaluating other elements

playing a role in such a relevant molecule. In this thesis, we have addressed this issue

through the study of an interesting cohort of 29 patients with moderate or severe

deficiency of antithrombin without mutation in the coding region of the SERPINC1

gene. The search also involved the analysis of 307 healthy controls, a family study

(GAIT) and large case control studies involving 2,980 patients and 3,996 controls from

different countries and diverse thrombotic diseases, as well as patients with congenital

disorder of glycosylation, which is a rare disorder. To achieve our objectives we have

used extensive molecular methodology ranging, from simple genotyping of point

mutations or polymorphisms to complex genome wide association studies, as well as

sequencing of two genes (SERPINC1 and PMM2) and quantitative RT-PCR. We also

used cellular biology methods, such as reporter systems, silencing assays, and

immunofluorescence staining. Finally, biochemical assays were also used, from western

blot analysis of multiple proteins (antithrombin, α1-antitrypsin, transferrin, prothrombin

and different platelet glycoproteins), to proteomic and glycomic studies. We also used

HPLC and FPLC methodology.

Our study has identified the first mutation affecting the promoter of SERPINC1

gene that is associated with antithrombin deficiency. This result has allowed us not only

to identify new regulatory regions of this gene, but also sustains the analysis of this

region in molecular studies of patients with severe or moderate antithrombin deficiency.

We also confirmed the low genetic variability of SERPINC1, particularly in the coding

region, probably explained by the conformational and functional sensitivity of this

serpin. Moreover, we found a minor role of SERPINC1 genetic variability on the levels

of this anticoagulant in healthy population. This result, together with the high

heritability of this trait, sustains the existence of other genetic factors modulating

antithrombin levels. Actually, our GWAS and validation studies together with further

functional results have identified the first modulating gene of antithrombin. Impaired

LARGE expression caused reduction of antithrombin, potentially by the intracellular

role of the heparan-like molecule generated by the xylosyltransferase and

Abstract

4

glucuronyltransferase activities of this molecule, although the mechanism is still

unknown. Our study has also identified global glycosylation abnormalities interfering

with antithrombin levels. Thus, we identified the first PMM2-CDG patient diagnosed on

his severe thrombotic history and deficiency of antithrombin. We also identified a new

disorder, called CDG-like, which shares the same biochemical pattern as PMM2-CDG

but only one clinical phenotype (thrombosis) and one out of the two required PMM2

mutation. Interestingly, these CDG-like disorders seems to have a relatively high

prevalence in patients with thrombosis. These results also suggest disorders of

glycosylation as strong thrombophilic disorders and encourage the analysis of this

disorder among patients with thrombosis and antithrombin deficiency but no mutations

in the SERPINC1 gene.

Our study also opens new attractive questions that must be answered: additional

key regulatory regions of the SERPINC1 gene must exist and they should be identified;

other antithrombin modulating genes must be explored and genetic modifications

affecting these genes should be studied in a context of a potential role in thrombosis; the

real incidence of CDG-like disorders on thrombotic diseases must be clarified; and the

mechanism involved in the abnormal haemostatic system and risk of thrombosis of

PMM2-CDG and CDG-like patients should be characterized.

Introduction

5

INTRODUCTION

Introduction

6

Introduction

7

1. Haemostatic system

During thousands of years of evolution, higher organisms with a closed and highly

pressured circulatory system have developed elements and mechanisms able to produce a

quick, simple and immediate response to the endothelial damage to minimize blood loss.

Most of these processes will not require protein synthesis to avoid the otherwise fatal

consequences of bleeding. The result is a finely regulated haemostatic system that includes

a network of interactions of vasoconstriction, platelet aggregation and coagulation

reactions. Finally, fibrinolysis slowly dissolves clots to maintain organ perfusion.

Physiological inhibitors in plasma maintain the balance between coagulation,

anticoagulation and fibrinolysis by localizing coagulation activity at the site of injury and

preventing systemic coagulation. In this introduction, we will briefly describe only

procoagulant reactions and their control, focusing on the main objective of our project:

antithrombin.

Exposure of tissue factor after vascular damage triggers serial proteolytic reactions in

cascade in which a zymogen (inactive enzyme precursor presented in plasma at a very high

concentration) of a serine protease and its co-factor are activated. These active proteases

then catalyze the next reaction in the cascade, ultimately resulting in thrombin formation,

the last procoagulant serine protease of the coagulation cascade. Thrombin generates a clot

of fibrin fibres by proteolyzing huge amounts of plasma fibrinogen. Thrombin also

activates FXIII, which by its transglutaminase activity cross-links fibrin fibres,

strengthening the clot. Finally, thrombin is also a potent platelet agonist. All these data

sustain the key role of thrombin in pro-coagulant and pro-aggregant reactions. As shown in

figure 1, other serine proteases are also crucial elements of the coagulation cascade

This strong procoagulant potential has to be precisely controlled in space and time:

procoagulant reactions must be restricted to the injured areas where & when the damage is

present. Otherwise, uncontrolled thrombin will lead to thrombosis by an obstructive

clotting disorder. For this reason, the fibrinolytic system, which is also based on serial

serine protease activation, slowly breaks down the clot formed (Figure 1). Finally, there are

other key proteins involved in the control of haemostasis: a myriad of inhibitors of

haemostatic proteases. According to their final function, these proteins can be classified in

a) inhibitors of fibrinolytic factors and b) natural anticoagulants [1].

Introduction

8

Figure 1. Clot cascade (grey), fibrinolysis (green) and anticoagulant molecules (yellow). Haemostatic

serpins are underlined. TM: thrombomodulin; PS: protein S; PC: protein C; PCI: protein C inhibitor; HCII:

heparin cofactor II; Fo: fibrinogen; ZPI: protein Z-dependent protease inhibitor; TFPI: Tissue factor pathway

inhibitor; TF: tissue factor; K: kininogen, PK: prekalikrein; uPA: urokinase-type Plasminogen Activator;

tPA: tissue plasminogen activator; PAI-I: plasminogen activator inhibitor; AT: antithrombin.

2. Thrombophilic factors

The haemostatic system necessitates a regulated equilibrium between strong

procoagulant elements triggered by vascular damage and even stronger anticoagulant and

fibrinolytic factors that maintain blood flowing. Accordingly, even minor disturbances of

potentially any of the elements of the haemostatic system can have pathological

consequences leading to an impair coagulation (hemorrhage), a hypercoagulation

(thrombosis), or paradoxically to both of them (disseminated intravascular coagulation).

The medical and socio-economical relevance of thrombotic disorders (the first cause of

morbid-mortality in the world) have encouraged the search for mechanisms or factors

increasing the risk to develop thromboembolic diseases.

The first description of a prothrombotic genetic defect arises in 1965, when the

Norwegian haematologist O. Egeberg studied a family with a great incidence of venous

thrombosis. Patients shared a common defect in the haemostatic system: deficiency of

TM FXa

FVa

PS

Fibrin

Plasmin

Fo

FXIII

FIIa TFPI

FXIa

TF

FII

Endothelium

Endothelium

t-PA

PAI-1

FIXa FVIIIa

FVIIa

FXIIa

K; PK, Kal

PC

AT HCII

ZPI

Plasminogen u-PA

PCI

Introduction

9

antithrombin [2]. A rare mutation affecting the gene encoding antithrombin explained both

the deficiency of this key endogenous anticoagulant and the high risk of thrombosis of

carriers due to the role of this anticoagulant. In the 80’s, the identification of other families

with thrombosis and deficiency of other anticoagulant proteins (protein C or protein S),

suggested that at least some cases of venous thrombosis might be considered as a

monogenic disease. However, the penetrance of these defects is variable. Moreover, the

incidence of all these defects collectively is low, even among thrombophilic patients (5-

10%) (Table 1). These data, together with the fact that more than 60% of the variation in

susceptibility to venous thrombosis is attributable to genetic factors [3], strongly suggested

the presence of additional prothrombotic genetic risk factors.

The search for genetic factors involved in thrombosis changed from rare mutations

associated with high risk to common polymorphisms following the attractive hypothesis of

“common genetic variants are involved in common diseases”. Certainly such interesting

hypothesis was validated in patients that had suffered thrombosis by the group of R.

Bertina, who in 1994 and 1996 identified two functional polymorphisms, the first

impairing the inactivation of FV (FV Leiden Arg506Gln) and the second associated with

moderately high levels of prothrombin (PT 20210G>A), which significantly increased the

risk of thrombosis [4] (Table 1). These reports clearly demonstrated that venous

thrombosis is a polygenic disorder and sustained the search for new prothrombotic

polymorphisms, particularly since heritable risk factors for venous thrombosis can be

identified only in 30–50% of affected patients [4]. This search was originally restricted to

candidate genes, mainly from the haemostatic system, using case-control studies. The

simplicity of such studies facilitated the sudden increase of thrombophilic studies (a

pubmed’s search using “polymorphisms and thrombosis” renders more than 2,000 entries).

However, the final conclusion of these studies is quite frustrating since most of them found

no significant association. Moreover, studies finding positive associations are usually

restricted to some specific group of patients or populations, or have not been validated by

other groups. Independently of this controversy, the risk of thrombosis associated with any

polymorphism is always moderate (OR<2). A recent meta-analysis including 126,525

cases and 184,068 controls from 173 case-control studies that had evaluated 28

polymorphisms from 21 genes found that only few polymorphisms mildly associated with

a thrombotic risk [5]. With the development of new genotyping and sequencing

technologies, the design of experiments to identify novel prothrombotic polymorphisms

changed, leading firstly to massive analysis of hundreds of polymorphisms located on

Introduction

10

candidate genes, many of them not restricted to the haemostatic system, to genome wide

association studies (GWAS) that genotyped up to 551,141 SNPs covering the whole

genome in thousands of cases and controls [6,7]. Unfortunately, these expensive studies

only confirmed classical associations and suggest that it is unlikely that new common

polymorphisms could play a key role in venous thrombosis. In contrast, recent excellent

studies identified new mutations with strong thrombophilic consequences affecting the

levels or function of different haemostatic proteins [8,9]. New thrombophilic defects must

be identified by a deep study of cases with idiopathic thrombosis.

Table 1. Frequent thrombophilic defects.

Genetic Defect Haemostatic consequence Prevalence

Patients

Prevalence

Controls

Relative

Risk (OR)

Antithrombin

deficiency

Reduced anticoagulant

capacity (50%)

0.5-2% 0.02% 10-50

Protein C

deficiency


capacity (50%)

3-5% 0.3% 10

Protein S

deficiency


capacity (50%)

2% 0.03-0.13% 8-10

Factor V

Leiden

Activated Protein C resistance 10-20% 4% 3-5

Prothrombin

20210G>A

High prothrombin levels 6-10% 3% 2.-3

ABO blood

group No-O

High FVIII&WF levels? 71% 50% 2

3. Serpins

As stated before, anticoagulation is crucial for maintaining blood flowing. If

procoagulant reactions are strong and efficient, anticoagulant mechanisms should be

superior. The primary objective of these proteins is the control of proteases unspecifically

generated in order to avoid spontaneous thrombosis. They should also avoid excessive or

mislocalized thrombin formation under vascular damage. This strong anticoagulant activity

is not too diverse. Although there are many proteins with capacity to inhibit procoagulant

proteases, there are three main natural anticoagulants: the tissue factor pathway inhibitor

(TFPI), the protein C, and particularly antithrombin. With minor relevance, other

anticoagulants are the protein Z-dependent protease inhibitor (ZPI), the heparin cofactor II

Introduction

11

(HCII), and the protein C inhibitor (PCI) (Figure 1). Other elements such as protein S,

thrombomodulin, endothelial receptor of protein C or different glycosaminoglycans will

also play a role in different anticoagulant pathways (protein C and antithrombin).

Interestingly, four of these anticoagulants (antithrombin, heparin cofactor II, the

protein Z-dependent protease inhibitor and protein C inhibitor) belong to the superfamily

of serpins. Actually, the features of these proteins make them ideal for the haemostatic

system, as we will discuss soon. This also explains the presence of additional haemostatic

serpins, this time controlling fibrinolytic reactions: α2-antiplasmin and plasminogen

inhibitor 1 (PAI-1).

The superfamily of serpins (SERine Protease Inhibitors) spans more than 1,500

proteins [10] due to the so close molecular, structural and functional similarities among

three proteins: human antithrombin, human α1-antitrypsin and chicken ovoalbumin [11].

The genes coding these proteins are divided into subgroups called clades (16 clades A-P)

based on various aspects of similarity [10]. The systematic name of each serpin is,

SERPINXy where X is the clade and y is the number within the clade [10]. Thus, the

appropriate names for the haemostatic serpins are the following: Protein C Inhibitor

(SERPINA5), Protein Z-dependent protease inhibitor (SERPINA10), heparin cofactor II

(SERPIND1), antithrombin (SERPINC1), α2-antiplasmin (SERPINF2) and plasminogen

inhibitor 1 or PAI-1 (SERPINE1).

Most of the serpins have serine protease inhibitor activity (although certain of them

have caspase activity or do not inhibit enzymes [12]) and this function makes them central

in the control of a big amount of proteolytic cascades inside or outside the cell.

Consequently, they play important roles in different processes such as inflammation,

apoptosis, complement, and as it can be expected coagulation and fibrinolysis [13-15].

Their sophisticate structure of serpins has played a central role in the understanding of their

function and biology. Most of the serpins comprise 400-500 residues, a molecular weight

of 37-70 kDa and an amino acid homology of 35% [16]. They are folded in a very

conserved tertiary structure with 3 β-sheets (A-C) and 8 or 9 α-helixes (A-I) with a flexible

and mobile region in the molecule pole responsible for the interaction with the target

protease, which is known as Reactive Centre Loop (RCL) (Figure 2).

Introduction

12

Figure 2. Native structure of α1-antitrypsin, the archetype structure of serpins (PDB: 1QLP). The

Reactive Centre Loop (RCL) is drawn in magenta, sheets A in blue, B in green, C in yellow and α helix (A-I)

in red. Dashed circles highlight the hinge position in the RCL and the shutter, both are areas very conserved.

Like other protease inhibitors, the RCL of serpins acts as a protease-specific bait.

However, unlike other families of serine protease inhibitors, which function in a simple

way by extending a substrate-like RCL that passively blocks the active centre of the target

protease [17], the RCL cleavage of serpins by target proteases results in inhibition by an

unique and extraordinary efficient suicide mechanism that has been compared with an

mouse-trap (Figure 3). This mechanism is possible because serpins do not fold to their

thermodynamically most favourable conformation. The serpin’s native state is metastable,

a kinetically trapped folding intermediate that upon RCL cleavage by proteases releases the

energy for a significant conformational transition resulting in the most stable cleaved

serpin [18]. The cleaved RCL moves to fill the middle strand of the main β-sheet of the

molecule (sA), and in doing so carries with it the entrapped protease. The protease is flung

a distance of 70 Å with an accompanying plucking and distortion of its structure that

ensures its destruction [19]. Moreover, covalent protease-serpin complexes are rapidly

cleared from circulation.

The conformational flexibility of serpins also makes them vulnerable to even minor

genetic or environmental changes, which may affect a correct folding and/or a correct

inhibitory function. These features, together with their relevance in important biological

regulatory processes make serpins prone to mutation-related diseases [20]. Mutations that

“Hinge”

“Shutter”

C Sheet

B Sheet

A Sheet

β Sheets

P1-P1´

Introduction

13

alter even a single amino acid in critical regions of these proteins can disrupt their function.

Abnormal serpins often form clumps (aggregates) that can build up to toxic levels within

cells. These protein aggregates may also cause a shortage of the inhibitor in areas where it

is needed to control chemical reactions. Disorders caused by aggregates of abnormal

serpins are called serpinopathies [20,21]. Mutations in SERPIN genes have been identified

as the cause of multiple disorders, including a lung disease called emphysema, hereditary

angioedema, a type of familial dementia, and abnormal blood clotting (thrombosis)

[10,22].

Figure 3. Inhibitory mechanism of Serpins.

4. Antithrombin

The first evidence of the existence of antithrombin can be found in 1982, when A.

Schmidt suggests the presence of a plasma cytoglobulin that prevented coagulation. The

name antithrombin was originally given by P. Moratowitz in 1905, while the first clinical

assay to determine the plasma levels of antithrombin was developed in 1963 by A. Hensen

y E.A. Loeliger. The first proof of the key role of antithrombin in haemostasis and

thrombosis was the association with venous thrombosis of a heterozygous deficiency

observed by O. Egeberg in 1965 [2].

Antithrombin is therefore an old well-characterised molecule. However, there are

still many fascinating points concerning this molecule.

Serpin Michaelis complex Final covalent complex serpin-protease

Protease

Introduction

14

4.1 Structure and biological function

Antithrombin is synthesized in the liver and released in plasma reaching an average

concentration of 150μg/ml. The molecule, with a plasma half-life of approximately 3 days,

is degraded by liver cells [23].

Antithrombin contains 432 residues accounting for a molecular mass of 58 kDa.

During its synthesis within the hepatocyte, three disulfide bridges are formed, a 32-

residues signal peptide is released by unknown proteases, and up to four glycosylation sites

are linked by the N-glycosylation machinery (Figure 4).

Figure 4. Localization of disulfide bridges and N-glycosylation sites in antithrombin.

The N-glycan present in antithrombin is biantenary and monosialylated, without

core-fucosylation, which may reduce the heparin affinity [24]. The exact composition and

linkages are shown in figure 5.

Figure 5. Composition of the N-glycans of antithrombin. NACGlu: N-Acetyl glucosamine; Glu:

glucose; Gal: galactose; Sial: Sialic acid.

The presence of a serine instead of threonine at position 137 reduces the efficiency of

N-glycosylation at the consensus sequence of Asn135 leading to the generation of two

Glycosylation sites

Processed polypeptide Signal peptide

Introduction

15

main glycoforms of antithrombin in plasma [25]: α-antithrombin, with four N-glycans,

represents up to 90% of total plasma antithrombin, and β-antithrombin lacking the N-

glycan at Asn135 (Figure 6). The different site-occupancy affects its size, pI, and more

structurally relevant its heparin affinity of the molecule, which is crucial for the activation

of this serpin [26]. The absence of the N-glycan at Asn135 increases 6-fold the affinity for

heparin [27]. Its higher heparin affinity sustains that despite of being the minority

antithrombin glycoform in plasma (10%), β-antithrombin could be the most relevant

thrombin regulator, particularly at the endothelial surface [28,29] and it could also explain

its increased clearance rate, which contributes to the smaller amount of β-antithrombin in

plasma [30]. Based on the clearance rate of the different antithrombin glycoforms, their

distribution between blood, non-circulating vessel surface and subendothelial

compartments, it has been proposed that α-antithrombin accounts for most of the

anticoagulant function in flowing blood, and that β-antithrombin plays a relatively greater

role at the vessel wall [31-35].

Figure 6. Main plasma glycoforms of antithrombin. The number and structural localization of N-

glycans are shown.

Introduction

16

This molecule shares the highly conserved structure of all serpins with three β-sheets

(A–C) and nine α-helices (A–I) [36,37] and the efficient and strong mechanism of

inhibition described for serpins, which for antithrombin is triggered after the proteolytic

cleavage of the bound between Arg-393 and Ser-394 residues (P1-P1’) by the target

proteases [38]. However, in contrast to other serpins, the RCL of native antithrombin

remains partially inserted into the top of the central A β-sheet (Figure 7) [39]. This native

conformation together with the saline bridge between P1 residue (Arg393) and Glu237 [40]

justify the relative inaccessibility of the target protease to the P1 residue and so the low

anticoagulant activity of the native form of antithrombin [41]. Actually, the full

anticoagulant activity of antithrombin requires the activation of this molecule triggered by

heparin and other glycosaminoglycans (Figure 7). In vivo, physiologically relevant forms

of heparin include heparan sulphate, found lining the endothelial layer and heparin released

from the granules of mast cells associated with the endothelium.

Figure 7. Native and active forms of antithrombin. The interactions of the P1 residue (Arg393) in the

native conformation are also shown.

The binding of heparinoid molecules to antithrombin occurs at the heparin binding

site of the serpin, which is centred on the D helix (Figure 2) [42]. An initial weak complex

with a specific pentasaccharide sequence in heparin is formed [43]. This causes complex

sequencial conformational changes in the molecule, where the D helix is extended closing

the central A-β and releasing the RCL. The activation of antithrombin culminates with the

Native Active

Introduction

17

final exposure of the P1 lateral chain adopting a more favourable conformation for binding

the target serine protease (Figure 8) [42,44,45].

Figure 8. Conformational changes on antithrombin induced by binding of heparin.

The conformational changes brought about by binding the pentasaccharide sequence

accelerates antithrombin’s interaction with a central coagulation enzyme, factor Xa (FXa),

by 100-fold [43]. Therefore, a synthetic form of this pentasacharide sequence is currently

used as an effective therapeutic anticoagulant an effective anticoagulant [46]. Moreover,

the exposition of this new exosites for FXa and FIXa contributes to increase the

anticoagulant activity of the active form [47].

Full acceleration of the inhibition of serine proteases, particularly that of another

major target, thrombin [43] but also FVIIa [48], requires the presence of heparin chains

which are longer than 26 residues [43] to facilitate the formation of a ternary complex by

serpin, protease and heparin (Figure 9) [45,49]. This complex accelerates the interaction

with serine proteases by up to 17,000-fold [43,45,49].

AT Native AT Active

50nM -25µM

Introduction

18

Figure 9. Inhibition of FXa and thrombin by antithrombin. Relevance of the length of heparin. AT:

Antithrombin. LMWH: Low molecular weight heparin.

The control of the activity of antithrombin by glycosaminoglycans has additional

effects, to delay and to localize this strong anticoagulant capacity. Otherwise, the high

levels of antithrombin in plasma might impair the procoagulant response necessary for a

correct haemostasis under vascular damage [50] (Figure 10).

Introduction

19

Figure 10. Localization of the anticoagulant activity of antithrombin (AT) to the endothelial surface by

heparan sulfates (HS). IIa: Thrombin.

Therefore, to properly fulfil its inhibitory mechanism, antithrombin needs an

extraordinary structural flexibility. In addition to the conformational changes shared with

other serpins, antithrombin also requires an additional conformational change to fully

activate it. Accordingly, this flexibility also makes antithrombin more susceptible than

other serpins to minor structural changes that could potentially result in a missfolded

conformation or inactive forms. Thus, in the normal native state, antithrombin contains a 5-

strand β−sheet (β−sheet A) and a surface-exposed RCL (bridging the C-terminus of strand

5A to the N-terminus of strand 1C). This state is, however, not the most stable. An

astounding increase in thermodynamic stability (best estimate – 32 kcal mol-1) [51] can be

achieved through the incorporation of the RCL into β-sheet A, triggered through extension

of strand 1C, to form the so-called latent conformation, which is not obviously inhibitory

(Figure 11). The identification of up to 5% of plasma antithrombin in a latent conformation

has suggested that this transformation could be a consequence of the senescence of the

native molecule [52,53].

Introduction

20

Figure 11. Native and latent conformations of antithrombin (PDB: 1T1FA and PDB: 2BEH,

respectively). The RCL is displayed in yellow, while the central β-sheet A is in red.

Finally, besides its key role in coagulation, antithrombin has also been described to

be important in other biological functions. A relaxed conformation seems to have anti-

angiogenic activity [54]. Moreover, a potent anti-inflammatory activity has also been

tightly associated with its heparin-binding domain [55]. Finally, our group has recently

found evidences suggesting a potential anti-apoptotic role of antithrombin [56].

4.2 Genetics

SERPINC1, the gene encoding antithrombin (GenBank X68793.1, OMIM# 107300)

is located at chromosome 1q23-25.1 and comprises 7 exons, encompassing 13.5 kb of

genomic DNA (Figure 12) [57,58] (start in 172.139.562 to 172.153.096 of pter; access

number GC01M172139: http://genecards.weizmann.ac.il/geneloc-

bin/display_map.pl?chr_nr=1&range_type=gc_id&gc_id=GC01M172139&contig=). It is

remarkable that only a few missense polymorphisms have been described, in concordance

with the structural and functional relevance of this molecule. So far, 267 polymorphisms

have been identified in SERPINC1, but its functional effects are unknown for most of them

(http://www.genecards.org/cgi-

bin/carddisp.pl?gene=SERPINC1&snp=267&search=serpinc1&rf=/home/genecards/curren

t/website/carddisp.pl#snp).

The promoter region of this gene is largely unknown. The promoter lacks TATA

element and GC-rich regions [58-60] and few studies, most of them using deletional

analysis and reporter or DNase I footprint assays, have identified few regions potentially

involved in the transcriptional regulation of this gene [60-63].

Latent Native

Introduction

21

Figure 12. SERPINC1 gen. Exons are marked as black boxes and Alu sequences with arrows.

4.3 Interindividual variability of antithrombin

Significant reduction of anticoagulant activity of antithrombin (values around 50%)

deeply increases thrombotic risk. However, as for many other haemostatic factors,

antithrombin levels have a wide range (70-120%) in general population. Factors such as

sex, body mass index, oral contraceptives or race, may influence antithrombin levels.

Furthermore, genetic factors must also play a role in determining antithrombin levels as

revealed by the high heritability of this trait (h= 0.486) [64]. Polymorphisms on the gene

encoding antithrombin (SERPINC1) could be involved in the interindividual variability of

antithrombin. Indeed, our group has described that a single nucleotide polymorphism

(SNP), rs2227589, located in intron 1 only explains up to 7.2% of the interindividual

variation of antithrombin [65]. Therefore, the search for new genetic factors associated

with the interindividual variability of antithrombin, with potential relevance on the risk of

thrombosis, must continue by evaluating other SERPINC1 polymorphisms, and other genes

that indirectly may influence plasma levels or the anticoagulant function of this key

haemostatic molecule. In fact, our group has previously confirmed the high conformational

and functional sensitivity of the molecule to different posttranslational modifications and

environmental factors [66-69].

4.4 Antithrombin deficiency and venous thrombosis

The extraordinary inhibitory mechanism of antithrombin and the large number of

procoagulant serine proteases able to be inhibited by this serpin (Figure 1) explain the key

haemostatic role of antithrombin. Therefore, even moderate reductions of antithrombin

activity increase the risk of venous thrombosis, as recently suggested [70]. As firstly

Introduction

22

described by Egeberg, heterozygous antithrombin deficiency significantly increases the

risk of venous thrombosis (OR: 10-50) and the complete absence of this molecule causes

embryonic lethality [71]. Moreover, the rate of thrombosis among carriers is significantly

high (increasing with age up to 60% in 60 year-old carriers) (Figure 13) as well as the risk

of recurrence [72].

Figure 13. Probability of thrombosis among carriers of different thrombophilic defects according to

age.

From a molecular point of view, more than 250 different mutations have been

identified in subjects with antithrombin deficiency (Bayston T, Lane D. Imperial College

of London. Antithrombin mutation database.

http://www1.imperial.ac.uk/departmentofmedicine/divisions/experimentalmedicine/haemat

ology/coag/antithrombin/ OMIM. http://omim.org/entry/107300 ; HGMD.

http://www.hgmd.cf.ac.uk/ac/gene.php?gene=SERPINC1). Two large and recent cohorts

have identified 69 new mutations [73,74]. Almost all antithrombin deficiencies are caused

by heterozygous mutations that for most cases are point mutations, although large gene

deletions have also been identified in SERPINC1. We have just described a relatively high

frequency of compound heterozygous carriers with antithrombin deficiency [75], all of

them carrying antithrombin Cambridge II, a common variant with a mild functional effect

[76]. Only few homozygous mutations have been described, all mild functional variants

[77].

Antithrombin deficiencies are normally classified attending to the presence of the

mutant protein in plasma. Thus, type I deficiency states are characterized by a complete

http://www1.imperial.ac.uk/departmentofmedicine/divisions/experimentalmedicine/haematology/coag/antithrombin/�


http://omim.org/entry/107300�

http://www.hgmd.cf.ac.uk/ac/gene.php?gene=SERPINC1�

Introduction

23

loss of the mutant antithrombin protein in plasma, resulting in the patient having 50% of

the normal levels and 50% of anticoagulant activity. This kind of antithrombin deficiency

is easily explained by non sense, frameshift, or splicing mutations, as well as large

deletions of the SERPINC1 gene. Interestingly, up to 26 missense mutations associate with

type I deficiency. Some of them have been explained by its effect on RNA stability.

Interestingly, other aminoacid changes, particularly those affecting mobile regions (mainly

the hinges of the RCL or the region involved in the shutter-like opening of the main β-

sheet), cause oligomer formation and intracellular retention, which has revealed a new

pathological mechanism causing antithrombin deficiency [78]. Thus, some cases of

thrombosis may be included within the group of conformational diseases [79]. Indeed,

antithrombin has been used to suggest a new and revolutionary mechanism of serpin

polymerization by domain swap [80]. Type I mutants are strongly correlated with a strong

predisposition towards venous thrombosis. In contrast, type II mutations are usually the

consequence of missense single amino acid changes that result in a protein which is

synthesized and usually secreted to the plasma with a normal or reduced rate, but they can

be detected at least by antigenic methods as the molecule is functionally defective.

According to the mutation-induced effect, three varieties of type II deficiencies have been

defined:

1. Heparin binding defect. These mutations impair the affinity of antithrombin

for heparin, resulting in a molecule that has normal inhibitory activity in the

absence of heparin, but reduced activity in its presence. Obviously, most mutations

causing this type of antithrombin deficiency affect directly the heparin biding

domain. However, other mechanisms might also cause heparin binding defects. In

this regard, our group has recently described an abnormal fucosylation of a variant

of antithrombin that impairs heparin affinity, resulting in a type II deficiency with

low heparin affinity [24], pointing out the relevance that glycosylation may have in

the function of this molecule.

2. Reactive site defect. This group includes mutations normally localized in the

RCL impairing the activity of the molecule independently of heparin binding.

3. Pleiotropic defect. This group includes a large variety of mutations with

multiple effects on the activation by heparin and the inhibitory function of the

protein.

The thrombotic risk associated with type II deficiencies is quite heterogeneous. Thus,

some cases have severe clinical consequences, while others have mild consequences as

Introduction

24

they have been identified in combination with other prothrombotic risk factors, such as

factor V Leiden or in homozygous or hemizygous states. Recent studies from our group

suggest different mechanisms to explain such a clinical heterogeneity. A post-translational

mosaicism may explain the mild thrombotic role of heterozygous mutations as well as the

survival of homozygous carriers of heparin binding defects, as the mutation does not

abolish the activation induced by heparin in the β-glycoform [77]. In contrast, other

mutations, in addition to a loss-of-function, also have a prothrombotic gain-of-function.

Thus, mutations affecting the P1 residue of the RCL (Arg393), particularly antithrombin

London (p.Arg393del), have higher heparin affinity that impairs activation of the wild-type

molecule with low concentrations of this glycosaminoglycan [81]. We have recently

described another gain-of-function for conformational mutations. The growing polymers of

conformationally unstable antithrombin mutants may incorporate wild-type monomers

under moderate stress conditions [82].

Accordingly, antithrombin assays are included among thrombophilic tests aiming to

identify congenital antithrombin deficiency. The challenge in managing these patients is

preventing potentially life-threatening thrombosis, while minimizing the equally

significant risk of haemorrhage associated with long-term anticoagulation. This is achieved

in the first instance by identifying high-risk episodes such as surgery, immobility and

pregnancy for which prophylactic anticoagulation can be used in the short term.

Prophylaxis for such periods is best provided by the use of low molecular weight heparin

(LMWH) with substitution or addition of antithrombin concentrate in particularly high-risk

circumstances. In the case of pregnancy, antithrombin concentrate is often used around the

time of birth when LMWH may increase the risk of post-partum hemorrhage. As patients

with congenital antithrombin deficiency get older, their thrombotic risk gradually increases

and for many patients long-term oral anticoagulant therapy becomes unavoidable because

of recurrent episodes of venous thromboembolism. The use of heparins might not be

recommended in patients carrying type II deficiency with heparin binding defect,

particularly in homozygous state. Concentrates of antithrombin for replacement therapy

have been classically obtained from human plasma, and there is now a recombinant

antithrombin with β-like properties [83].

Introduction

25

Objectives

27

OBJECTIVES

Objectives

28

Objectives

29

Hypothesis

There might be unknown elements modifying antithrombin levels with potential

thrombotic implications as the current known elements do not explain themselves either

this wide variability or the high heritability of this trait in the population and a

significant proportion of patients with antithrombin deficiency without mutations in the

SERPINC1 gene.

The main objective of this thesis was to identify new genetic elements playing a

role on the levels or function of plasma antithrombin and to determine possible

physiopathological implications.

Specific objectives:

1. To identify genetic alterations in SERPINC1 gene associated with

interindividual variability of antithrombin and/or thrombotic risk.

2. To find out new antithrombin modifier genes and the potential molecular

mechanisms underlying.

3. To assess the importance of a correct N-glycosylation on antithrombin levels and

on the control of normal haemostasis.

Objectives

30

Material and methods

31

MATERIAL

AND METHODS


32


33

1. Patients and healthy subjects

This study included the following cohorts of subjects:

1) Three hundred and seven Spanish Caucasian healthy blood donors (138 male/169

female), with an average age of 43 years.

2) One hundred and eighteen subjects with antithrombin deficiency from 22 Spanish

hospitals, 1 Portuguese hospital and 1 Israeli hospital. Two different groups were

selected from this cohort.

a. Twenty-nine unrelated thrombophilic patients with moderate antithrombin

deficiency (70-90% anti-FXa activity of the reference value). In this group,

sequencing analysis of SERPINC1 revealed a relatively common point

mutation in exon 7 responsible for the Cambridge II variant (p.Ala384Ser)

[76] in four patients. Thus, our study was mainly focused to identify new

factors that might be involved in the reduced antithrombin levels of the

remaining 25 patients (named as P1-P25).

b. Eighty nine unrelated and consecutive subjects with verified antithrombin

deficiency (30-70%). Molecular analysis identified a genetic defect in the

SERPINC1 gene in 85 cases that explained the associated deficiency.

Therefore, we addressed our interest to identify the mechanism underlying

antithrombin deficiency of the remaining 4 cases that had neither gross

deletion, nor mutation in the coding or flanking regions of the SERPINC1

gene. In this rare but scientifically-fascinating group of subjects, we

differentiated two subgroups based on familial history

i. Two subjects (P26 and P27) that reported no relatives with

antithrombin deficiency.

ii. Two subjects that had up to three relatives from two generations with

type I antithrombin deficiency (F1 and F2), sustaining a congenital

defect.

3) Two patients with congenital disorder of glycosylation. Prof. J Jaeken kindly

supplied two young Belgian patients with congenital disorder of glycosylation

(CDG) type Ia, also named PMM2-CDG. The first patient carried two mutations in

PMM2 gene responsible for the amino acidic changes p.Phe119Leu and


34

p.Arg141His. The second patient also carried two mutations in PMM2 leading to

p.Pro113Leu and p.Thr237Arg substitutions, respectively.

Informed consent was obtained from all donors, patients or their legal representatives.

2. Blood sampling

Blood samples were obtained by venopuncture collection into 1:10 volume of

trisodium citrate. Platelet poor plasma fractions were obtained (within 5 min after blood

collection) by centrifugation at 4 ºC for 20 min at 2200xg and stored at -70 ºC. Genomic

DNA was purified by the salting out procedure and stored at -20 ºC.

All subjects gave their informed consent to enter the study, which was approved by the

ethics committee of the Centro Regional de Hemodonación (Murcia, Spain) and performed

according to the declaration of Helsinki, as amended in Edinburgh in 2000.

3. Antithrombin levels and other thrombophilic tests

Routine thrombophilic tests including protein C, protein S, antiphospholipid

antibodies, FV Leiden, and prothrombin G20210A were performed, as described elsewhere

[84]. Plasma FXa-inhibiting activity was measured using a chromogenic method in presence

of heparin (HemosIL TH, Instrumentation Laboratory) as previously reported [68].

Antithrombin antigen levels in plasma were determined by rocket immunoelectrophoresis

and/or a home-made ELISA. For both parameters, values were expressed as a percentage of

the result observed in a pool of citrated plasma from 100 healthy subjects.

4. Amplification and sequencing of SERPINC1

Polymerase chain reactions (PCR) covering the whole SERPINC1 gene and the

potential promoter region were carried out using Expand Long Template Polymerase

(Roche, Spain) and the oligonucleotide sets described in Table 2. PCR products were

purified and sequenced with ABI Prism Big Dye Terminator v3.1 Cycle sequencing kit and

resolved on a 3130xl Genetic Analyzer (Applied Biosystems, Spain). Comparison with the

reference sequence (GenBank accession number NG_012462.1) was performed with

SeqScape v2.5 software (Applied Biosystems, Spain).


35

Table 2. Oligonucleotides used for PCR amplification of the SERPINC1 gene. E: Exon; I: Intron. F:

forward primer; R: reverse primer.

Amplicon Primer name Oligonucleotide sequence (5´-3’) Size (bp) 5’ region AT_-1F GGACCTTATTAACATCTGAC

743 AT_-728R CCAAGTTCACAGGGCGCATG

AT_-660F GGAGTCCTTGATCACACAGCA 1088

AT_IN1R GTCTTTGACTGTAACTACCAG

AT_0,5F GGCTACTGTTCACACTGCACATG

1067 AT_1R GAAAGCTCACCCCTCTTAC

E1-E2 AT_1F CTGTCCTCTGGAACCTCTGCGAGA

2880 AT_2R GGTTGAGGAATCATTGGACTTG

I1 AT_19F ATCCGGGAAGAGAGCAAATGC

1200 AT_31R CTCAAAACAGCAACAACAAAC

E2-E3 AT-2F TGCAGCCTAGCTTAACTTGGCATTT

3180 AT_3aR AGAGGAAGAACTCGGAGGTCAGG

E3-E4 AT_3aF AACTAGGCAGCCCACCAAACCC

1410 AT_3bR GAAGAGCAAGAGGAAGTCCCT

E4-E5 AT_3bF TTGAATAGCACAGGTGAGTAGGTT

1380 AT_4R AAGGGAGGAAACTCCTTCCTAG

E5 AT_4F TGTGTTCTTACTTTGTGATTCTCT

391 AT_4R AAGGGAGGAAACTCCTTCCTAG

E5-E6 AT_4F TGTGTTCTTACTTTGTGATTCTCT

2521 AT_5R CATGCATCTCCTTTCTGTACCC

I5-E6 AT_95F TGGAGAGGAATTTGAAAG

1800 AT_5R CATGCATCTCCTTTCTGTACCC

I5 AT_95F TGGAGAGGAATTTGAAAG

1080 AT_106R AACATCTTTCTTTCCAGTCTG

E6-E7 AT_5F TTCTCCCATCTCACAAAGAC

3660 AT_6R AGAGGTGCAAAGAATAAGAA

I6 AT_130F GGGAGACTAGGGTGTTGA

1200 AT_142R CAGATCTAGAGGGAAACACC

I6-E7 AT_130F GGGAGACTAGGGTGTTGA

1860 AT_6R AGAGGTGCAAAGAATAAGAA

3’ region

AT_130F GGGAGACTAGGGTGTTGA 2860

AT_159R ACACACAAGTAATAACATCCAC

AT_3’F TTATCTTCATGGGCAGAGTAGC

1000 AT_159R ACACACAAGTAATAACATCCAC

5. Multiplex Ligation-dependent Probe Amplification (MLPA)

Multiplex Ligation-dependent Probe Amplification technique, a multiplex PCR

method able to distinguish sequences differing in only one nucleotide, was used to detect

abnormal copy numbers of up to 50 different genomic DNA or RNA sequences. To


36

specifically detect deletions or duplications in the SERPINC1 gene we used the SALSA®

MLPA® Kit P227 SerpinC1 (MRC-Holland).

6. Genotyping of SERPINC1 genetic alterations

The genotype of rs5778785 was determined by high resolution melting (HRM)

analysis following the Type-it HRM PCR protocol (Qiagen, Spain). The g.2085 T>C

transition was genotyped by a HRM/razor probe technique (Figure 14).

Figure 14. HRM-razor melting curves and electropherograms of wild type and heterozygous g.2085 T>C

genotypes.

These two HRM analyses were carried out in Rotor-Gene Q 6000 (Qiagen, Spain)

with the primers and probes shown in Table 3.

Nine SERPINC1 polymorphisms were genotyped by PCR-allelic specific restriction

assays using primers and restriction enzymes described in Table 3, as previously described

[85]. For some polymorphisms, mutated primers were designed (Table 3).

rs2227589 was genotyped by a Taqman SNP Genotyping Assay (C__16180170_20,

Applied-Biosystem), as previously described [65].

The g.2143 C>G mutation was studied by PCR-allelic specific restriction assay using:

5’-GGAGTCCTTGATCACACAGCA-3’ forward and 5’-

GTCTTTGACTGTAACTACCAG-3’ reverse primers and Bsa JI restriction enzyme (New

England BioLabs, Spain) following manufacturer’s specifications.

For all genotyping methods, results were confirmed by sequencing.

TºC


37

Table 3. Genotyping by PCR-allele specific restriction assay and High Resolution Melting. Mutated nucleotides are underlined.

Polymorphism Amplification/Probe Oligonucleotides PCR Size (bp) Method of genotyping

rs2227588 AT_0.5F: GGCTACTGTTCACACTGCACATG AT_-728R: CCAAGTTCACAGGGCGCATG 198 Mnl I

(NEB)

rs2227590 AT_18F:CATTTCAAGTGCTCTCCCTCC

203 FokI (Takara) AT_19R:GCATTTGCTCTCTTCCCCGAT

rs2227596 AT_30F:ATCTTCAGCACACAGATATTG

229 Tsp 45I (NEB) AT_33R:GCTGGGTGTGGTGGCGC

rs2227603 AT_53F:TTGAGACCAGGGGCCCAG

408 FokI (Takara) AT_57R:AATTGACCACTAGAATAGAAG

rs941989 AT_60F:GCTTCCAAACCACACATGTTC

241 HinfI (Promega) AT_62R:CCATGCTCTCTCCTCAGGC

rs2227607 AT_71F:GACTGCTGGAGGGCTGAC

186 NlaIII (NEB) AT_73R:GAACTGGCATCTTCAAAACACCC

rs2759328 AT_112F:TCCAAACTGAATTCCCATCTGCGG

190 FauI (Fermentas) AT_5R: CATGCATCTCCTTTCTGTACCC

rs677 AT_114F:TGAGTACACCTTCCCCAC

126 DdeI (NEB) AT_115R:CTGTATTATAGCAGGCCTGTGG

rs2227616 AT_130F:GGGAGACTAGGGTGTTGA

332 BstEII (NEB) AT_133R:CCGGGCACGGTGGCTCATGCCGG

TAA

rs5778785 AT_95F:TGGAGAGGAATTTGAAAG

178 HRM AT_97R:GTGAGCCGAGATCGCACCGC

g.2085 T>C (rs78972925)

AT_13F:GAGGGAGTGTGGGCAAGAGA 107

HRM/Razor probe AT_13R:AGGGTTAAGCAAAGTGTAGAG

Razor:ACCCGGACTAACTTGAAAT Razor probe

NEB: New England BioLabs; HRM: High Resolution Melting

7. Prediction of transcription factor binding sites to the SERPINC1 potential

promoter

The search for sequences with potential transcriptional relevance affected by genetic

variations identified in this study was performed by TESS and TFSEARCH programs, based


38

on TRANSFAC databases (http://www.cbil.upenn.edu/cgi-bin/tess/tess and

http://www.cbrc.jp/research/db/TFSEARCH).

8. Cloning and reporter assay

The 5’ non transcribed portion of SERPINC1 (838 bp) from one individual

heterozygous for the g.2143 C>G mutation was amplified using Expand High Fidelity DNA

polymerase (Roche, Spain) to minimize nucleotides misincorporation with 5’-

GGAGTCCTTGATCACACAGCA-3’ (forward) and 5’-AATCTCGCAGAGGTTCCAGA-

3’ (reverse) primers. PCR products were ligated into a T/A cloning vector (pCR®2.1,

Invitrogen, Spain). Clones with wild-type and mutant alleles were selected by PCR-allelic

restriction assay with Bsa JI, and the inserts cloned into Kpn I and Sac I sites of luciferase

pGL3 basic vector (Promega, Spain). Competent cells were transformed, plasmid DNA was

purified using a commercial kit (Plasmid Midi Kit, Qiagen, Spain), and they were sequenced

in both directions to faithfully confirm the wild-type and mutant sequences.

The ability of each sequence to promote transcription of luciferase gene was

transiently tested in human cell lines PLC-PRF-5 (human hepatome cell line) and HepG2

(Human hepatocellular liver carcinoma cell line) both with antithrombin constitutive

expression.

All plasmid DNA were quantified by Nanodrop (Thermo Scientific, Spain). Eighteen

hours before transfection, cells were plated at 80,000 per well of a 24-well plated (3

replicates per clone for each cell line and in each experiment) and cultured in Dulbecco's

Modified Eagle Medium (Invitrogen, Spain), at 37 ºC with 5% CO2. Reporter plasmid

(1,000 ng) along with 100 ng of Renilla luciferase control plasmid pRL-TK (Promega,

Spain) were transfected using Lipofectamine 2,000 (Invitrogen, Spain) for PLC-PRF-5 cell

line and GenJet reagent (SignaGen® Laboratories) for HepG2 cell line, according to the

manufacturer’s instructions. Luciferase assays were performed 48 h after transfection by

using the Dual-GloTM Luciferase assay System (Promega, Spain). Promoter activity was

normalized by dividing luciferase activity by Renilla luciferase activity for each transfected

well. pGL3 control vector (Promega, Spain) was used as a transfection positive control.

9. Genome wide association study

We carried out a genome wide genotype-phenotype association study in the GAIT

(Genetic Analysis of Idiopathic Thrombophilia) study, which included 352 individuals from

http://www.cbrc.jp/research/db/TFSEARCH�


39

21 extended Spanish families [3]. Twelve of these families were selected on the basis of a

proband with idiopathic thrombophilia, whereas the remaining 9 families were selected

randomly. Average pedigree size was 19. Importantly, no family had congenital


A genome-wide set of 307,984 SNPs was typed in all of the participants using the

Infinium® 317k Beadchip on the Illumina platform (San Diego, CA, USA). Genotype

imputation was performed with Merlin [86] to avoid missing values and all genotypes were

checked for Mendelian inconsistencies. In addition, any SNP with call rate<95%,

MAF<0.025 or failing to fit Hardy-Weinberg proportions taking into account multiple

testing (p<5×10-7) was removed from the study. In total, 24,547 SNPs failed to pass the data

cleaning criteria, leaving a set of 283,437 SNPs for further analysis.

10. Validation study

A cohort of 307 Spanish Caucasian healthy blood donors (138 males/169 females)

with a mean age of 43 years and from a different region than the GAIT study was selected to

replicate significant associations identified in the GWAS. Genotyping was performed using

TaqMan® probes specified in Table 4.

Table 4. TaqMan® probes used for genotyping in the validation study. The SNPs identified in the GWAS

associated with anti-FXa activity and the genes affected are also shown. *rs713703 located in LARGE and

strongly linked to rs762057, was not evaluated.

TaqMan® probe SNP Gene

c_15834925_20 rs2152192 SAMD3

c_2747618_20 rs1411771 DISC1

c_9839576_10 rs13193455 LOC 154449

c_1282266_10 rs11681944 C1D

c_1298059_10 rs6768189 CACNA2D3

c_ 30869706_10 rs10880942 SLC38A1

c_11634765_10 rs1860867 COBL

c_16207835_20 rs2356895 LOC283553

c_30230995_10 rs9896932 CD7


40

TaqMan® probe SNP Gene

c_2486441_10 rs762057 LARGE

c_1839936_10 rs713703 LARGE

c_790509_10 rs240082 LARGE

11. Case-control study

The SNP found to be significantly associated with antithrombin levels in the GWAS

and verified in the replication study, rs762057, was genotyped using a TaqMan® probe

(c_2486441_10) (Applied Biosystems) in a case-control study including 1018 patients with

venous thrombosis and 1018 controls previously described [87]. Briefly, patients with

venous thrombosis diagnosed appropriately were enrolled from the files of the

anticoagulation clinics in four Spanish Hospitals. Patients with known malignant disorders

were excluded. Controls were randomly selected among two sources: blood donors and

traumatology and ophthalmology patients matched by age, gender, race, and geographical

distribution with 1018 cases.

12. Statistical analyses

In all studies, deviation from Hardy-Weinberg equilibrium (HWE) was investigated

using a standard χ2 with 1 degree of freedom. Allele and genotype frequencies, haplotype

analysis, association of haplotypes with anti-FXa activity and linkage disequilibrium

analysis were calculated with the SNPstats software [88].

In the GAIT study, HWE was tested using parental data only. Association between

SNPs and plasma anti-FXa activity was tested using a measured genotype association

analysis assuming additive allele effects. This analysis was carried out using the variance-

components methodology implemented in the SOLAR Version 4.0 software (Southwest

Foundation for Biomedical Research, http://solar.sfbrgenetics.org/download.html) [89].

Analyses were adjusted for age, gender, and body mass index, and in females, for oral

contraception as well.

Association between SNPs and plasma anti-FXa activity in the validation study was

tested by a Student's t-test following a dominant model and adjusted for factors described to

influence antithrombin levels (age, gender and the SERPINC1 rs2227589 polymorphism)


41

[65,90]. Data are presented as mean ± standard deviation. This analysis was carried out

using the Statistical Package for Social Science (SPSS version 15.0, USA).

In the case-control study, univariate and multivariate analysis were performed using

the SPSS statistical package.

13. LARGE and SERPINC1 gene expression

LARGE gene expression was assessed in mononuclear cells of 10 healthy subjects by

qRT-PCR using Hs00893935_m1 TaqMan® Gene Expression Assay (Applied Biosystem)

and beta-actin (Hs99999903_m1) as constitutive reference gene. The correlation between

plasma anti-FXa activity and LARGE expression was determined by linear regression. These

analyses were also done in HepG2 and Human Embryonic Kidney cells expressing the

Epstein Barr Nuclear Antigen 1 (HEK-EBNA) cell lines transfected with LARGE gene

silencers.

SERPINC1 gene expression in HepG2 and HEK-EBNA cell lines transfected with

LARGE gene silencers was determined by qRT-PCR with SYBR® Green-Based Detection

(Applied Biosystem) using Tubuline beta-2C chain as constitutive reference gene. Primers

for amplification were: SERPINC1-F: TGTTCAGCCCTGAAAAGTCC; SERPINC1-R:

GCTGCTTCACTGCCTTCTTC; TUBULINE-F: GGCCGGACAACTTCGTTT and

TUBULINE-R: CCGCGCCTTCTGTGTAGT.

14. Purification of plasma antithrombin. Proteomic and glycomic analysis

Plasma predominant antithrombin glycoform (α) from two subjects selected from 10

healthy blood donors, with the highest and lowest LARGE expression values, as well as

antithrombin minor glycoform (β, with 3 N-glycans) from a pool of 100 healthy blood

donors were purified by Fast Protein Liquid Chromatography (FPLC) (ÄKTA Purifier from

GE Healthcare) using heparin affinity chromatography on HiTrap Heparin columns (GE

Healthcare), in 100 mM Tris-HCl and 10 mM citric acid, with a gradient from 0 to 0.25M

NaCl and a step of 2M NaCl. Fractions with antithrombin were applied to a HiTrap Q

column (GE Healthcare). Finally, proteins eluted were desalted through a dialysis tubing

(Sigma Aldrich) and stored at -70 ºC, prior to analysis. The molecular mass and glucidic

components of these molecules were determined by MALDI-TOF-MS (Matrix Assited

Laser Desorption Ionization-Time-of-Flight) analysis and HILIC (Hydrophilic Interaction

Liquid Chromatography) HPLC (High Performance Liquid Chromatography). Briefly, for


42

protein molecular weight determination a solution of 3,5-dimethoxy-4-hydroxycinnamic

acid (10 g/L) in acetonitrile (ACN)/water/trifluoroacetic acid (TFA) (50:50:0.1 by vol.) was

employed. Experiments were carried out on a Voyager-DE™ STR Biospectrometry

workstation (Applied Biosystems), equipped with a N2 laser (337 nm). Samples were

measured both in the linear, providing information on the total number of different

structures, and in the reflectron mode for identification of molecular formulas based on

precise mass measurements. Recorded data were processed with Data Explorer™ Software

(Applied Biosystems). The analysis of the N-glycans was performed by HILIC

chromatography. Briefly, N-glycans were released with N-glycosidase F (Roche Diagnostics

GmbH, Mannheim) following prior denaturing (5 min at 95 ºC in 150 mM sodium

phosphate buffer, pH 7.4). Afterwards, samples were chilled on ice and digested with 0.6 U

N-glycosidase F by incubation at 37 ºC, for 15 hours. Glycans were labeled as described

[91] and subjected to chromatographic separation on an Agilent 1100 HPLC equipped with a

fluorescence detector (1100 Agilent fluorescence module) using excitation and emission

wavelengths of λ=330 nm and λ=420 nm, respectively. The following gradient conditions

were employed on a ACQUITY UPLC™ BEH HILIC column (2.1 x 150 mm, 1.7 µm):

solvent A was 10% 50 mM ammonium formate (pH 4.4) in 90% ACN, solvent B was 90%

50 mM ammonium formate (pH 4.4) in 10% ACN, and the flow rate was 15 µl/min.

Following injection, samples were eluted by a linear gradient of 20-55% B over 100 min,

followed by a linear gradient of 55-100% B over the next 5 min. The column was eluted

using 100% B for 2 min, and subsequently re-equilibrated in 20% B before injection of the

next sample. The system was calibrated in glucose units (GU) using a 2-AB-labelled dextran

hydrolysate. The total running time was 125 min [91]. Mass spectrometric analyses of 2-

aminobenzamide (2-AB)-labeled glycans were performed 2,5-dihydroxybenzoic acid (DHB)

matrix (10 mg/ml) in ACN: H2O (50:50 v/v). Typically, spectra of sialylated N-glycans were

acquired in linear mode with negative polarity, and in neutral N-glycans reflectron mode and

positive polarity. External calibration of the spectrometer was performed using a mixture of

2-AB-labelled glucose oligomers in the positive-ion mode and 2-AB-derivatised fetuin N-

glycans in the negative mode. Recorded data were processed with Data ExplorerTM

Software (Applied Biosystems).


43

15. LARGE gene silencing and effect on N-glycoproteins

For these experiments we used two cell lines expressing antithrombin: HepG2 with

constitutive antithrombin expression, and HEK-EBNA transiently transfected with pCEP4-

AT plasmid (generously provided by Prof. JA Huntington) that expressed high levels of the

beta glycoform of human antithrombin [92].

HepG2 and HEK-EBNA cell lines were grown to 60% confluence at 37 °C, 5% CO2,

in DMEM (Invitrogen) supplemented with 5% fetal bovine serum (Sigma-Aldrich). Then,

they were transfected with 50 nM of specific LARGE siRNA: s17620 (Applied Biosystems)

for 30 minutes in OptiMEM with siPORT™ (Applied Biosystem). Appropriate controls:

transfections without siRNA, or with 50 nM of scramble siRNA (Silencer® Negative

Control AM4611, Applied Biosystem) were used. After 12 hours, cells were washed with

PBS and exchanged into CD-CHO medium (Invitrogen) supplemented with 4 mM L-

glutamine (Invitrogen). Cells were grown at 37 ºC for 48 hours. Then, RNA was purified

using TRIzol® Reagent (Invitrogen) following manufacturer instructions. We determined

the silencing efficiency evaluating LARGE and SERPINC1 expression by qRT-PCR, as

indicated above. Additionally, conditioned medium was harvested and in case of HepG2 cell

cultures, concentrated 5-fold using a CentriVap Concentrator (Labconco). The levels of

secreted antithrombin, transferrin, prothrombin and α1-antitrypsin in conditioned medium

were determined by western blotting, essentially as described elsewhere [93]. Briefly,

electrophoresis was carried out using sodium dodecyl sulfate–polyacrylamide gel

electrophoresis (SDS-PAGE) in 10% (w/v) polyacrylamide gels under reducing conditions.

Proteins were transblotted onto a polyvinylidene difluoride membrane. Proteins were

immunostained with specific rabbit [anti-human antithrombin (Sigma Aldrich) and anti-

human α1-antitrypsin (Dako Diagnostics, Denmark)], goat [anti-human transferrin (Sigma

Aldrich)], or sheep [anti-human prothrombin (Cerdalane laboratories, Canada)] polyclonal

antibodies; followed by proper secondary IgG-horseradish peroxidase conjugates (GE

Healthcare), and ECL detection (GE Healthcare). Antithrombin levels in the conditioned

medium were also determined by ELISA, as previously described [93]. Additionally, anti-

FXa activity of conditioned medium was measured by the chromogenic method described

above. Finally, we also evaluated the intracellular content of antithrombin by western

blotting and immunofluorescence, basically as previously described [93]. Briefly, cells were

extensively washed with sterile PBS and then lysated with 50 μl of lysis buffer (10 mM

TrisHCl, 0.5 mM DTT, 0.035% SDS, 1 mM EGTA, 50 mM sodium fluoride, 50 μM sodium


44

orthovanadate, 5 mM benzamidine and 20 mM phenylmethylsulphonyl fluoride) and stored

at -70 ºC, prior to analysis. Intracellular antithrombin was evaluated by Western blotting,

essentially as indicated above. For immunofluorescence analysis, cells were fixed with an

equal volume of 4% paraformaldehyde in PBS buffer pH 7.4 (22 ºC, 20 min). After fixation,

cells were washed with PBS, permeabilized with 0.1% Saponin, 0.2% Gelatin, 0.02% Azide

(3x5 min). All subsequent incubations and washes contained 0.1% Saponin, 0.2% Gelatin,

0.02% Azide in PBS buffer. Anti-antithrombin antibody was used at 1:1,000 and incubated

for 1 h at 22 ºC. Indirect immunofluorescence was carried out using the appropriate

fluorescein conjugated goat anti-Rabbit IgG (Vector laboratories, USA) 1:1,000.

Fluorescence was analyzed on a Confocal Microscope LEICA TCS-SP2 using its associated

software (Leica Microsystems).

16. Electrophoretic evaluation

Electrophoretic analyses were performed by using SDS and native PAGE under

reducing and non reducing conditions as previously reported [94]. Plasma samples were

transblotted onto a PVDF membrane and immunostained with different antibodies: rabbit

anti-human antithrombin polyclonal antibody (A9522, Sigma-Adrich), rabbit anti-human

α1-antitrypsin polyclonal antibody (A0012, Sigma-Aldrich), goat anti-transferrin polyclonal

antibody (T2027, Sigma-Aldrich), followed by donkey anti-rabbit IgG-horseradish

peroxidase conjugate (NA9340V, GE Healthcare) or mouse monoclonal anti-Goat/Sheep

IgG–Peroxidase (A9452, Sigma-Aldrich) with detection via an EC kit (Amersham

Biosciences, Piscataway, NJ).

17. N-glycosidase F, neuraminidase and galactosidase plasma treatment

Plasma from probands and healthy donors (1µl) were treated with 1 U α2–3,6,8,9

neuraminidase (sialidase) (N 3786, Sigma-Aldrich, Saint Louis, USA) at 37 ºC for 24h in 50

mM sodium phosphate buffer, pH 6. Afterward, samples were treated with β-galactosidase

(G 1288 Sigma-Aldrich, Saint Louis, USA) at 37 ºC for 24h in the same buffer.

Treatment with N-glycosidase F (Roche Diagnostics GmbH, Mannheim, Germany)

required a previous denaturing step (5 min at 95 ºC in 150 mM sodium phosphate buffer, pH

7.4). Then, samples were chilled on ice and digested with 0.6 U N-glycosidase F by

incubation at 37 ºC for 24h.


45

18. Purification and proteomic analysis of an antithrombin variant from a

CDGIa-like patient

Antithrombin variant from the patient was purified by heparin affinity chromatography

at pH 7.4 onto HiTrap Heparin columns (GE Healthcare, Barcelona, Spain), as previously

described [24] using an ÄKTA Purifier (GE Healthcare, Barcelona, Spain) in 50 mM Tris-

HCl, in a gradient from 0.15 to 2 M NaCl. Fractions with variant antithrombin were applied

to a HiTrap Q column (1 mL) (GE Healthcare, Barcelona, Spain). Proteins eluted with 50

mM Tris-HCl pH 7.4, in a gradient from 0 to 1 M NaCl. Fractions containing variant

antithrombin were pooled, desalted over 5 mL HiTrap Desalting columns (GE Healthcare,

Barcelona, Spain) and stored at -70ºC. Protein purity was checked by 8% SDS-PAGE.

A solution of 3,5-dimethoxy-4-hydroxycinnamic acid (10 g/L) in acetonitrile

(ACN)/water/trifluoroacetic acid (TFA) (50:50:0.1 by vol.) was chosen for protein analysis.

Experiments were carried out on a Voyager-DE™ STR Biospectrometry workstation

(Applied Biosystems, Madrid, Spain), equipped with a N2 laser (337 nm). Recorded data

were processed with Data Explorer™ Software (Applied Biosystems, Madrid, Spain).

19. Identification and quantification of transferrin glycoforms

Transferrin in serum was completely saturated with iron by mixing 100 µL of serum

with 20 µL of FeNTA (final concentration, 1.67 mM), a well-known transferrin iron donor.

Each transferrin molecule can bind a maximum of two iron ions, but typically, serum

transferrin is only partially (~30%) saturated. Because the FeNTA complex gives almost

instant iron saturation of transferrin, no additional incubation time was necessary at this

stage. Thereafter, the lipoproteins in the serum sample were precipitated by addition of 20

µL of the dextran sulfate–CaCl2 solution (10% dextran sulfate, 1 mM CaCl2). After gentle

mixing, the samples were left in the cold (4 °C) for 30–60 min and then centrifuged at 3500g

at 4 °C for 5 min. Of the clear supernatant, 130 µL were withdrawn and concentrated at

3500g at 4 °C in an YM-10 Microcon filter (Millipore, Darmstadt, Germany). The

concentrated sample was then diluted with 260 µL of water and transferred to glass HPLC

vials.

We injected 100 µL of the treated serum samples into the HPLC system, which

consisted of an Agilent 1100 Series Liquid Chromatography, equipped with a quaternary

pump and degasser, thermostated autosampler and column compartment, multiple

wavelength detector, and ChemStation software (Agilent Technologies, Madrid, Spain).


46

Separation of the transferrin glycoforms was performed on a SOURCE® 15Q PE 4.6/100

anion-exchange chromatography column (Amersham Biosciences) at 22 °C, by linear salt

gradient elution at a flow rate of 1.0 mL/min. Buffer A consisted of 10 mM Bis-Tris,

adjusted to pH 7.0 with 2 M HCl; buffer B was the same buffer containing 0.5 M NaCl, pH

6.2; and buffer C was 10 mM Bis-Tris, pH 6.2. After every sample injected, the column was

regenerated and cleaned by washing with 2.0 M NaCl (buffer D) and finally reequilibrated

with the starting buffer [95]. All buffers were degassed and filtered through a 0.45 µm filter

before use. Total run time under these conditions was 40 min. Quantification of the

transferrin glycoforms relied on the selective absorbance of the iron–transferrin complex at

470 nm referred as the area [95].

Q-TOF (Quadrupole Time-Of-Flight) analysis of transferrin forms was done by Dr. D

Lefeber (Netherlands) following a method developed in his laboratory (Monique Van

Scherpenzeel; Gerry Steenbergen; Dirk Lefeber N-glycan profiling of human serum

transferrin using microfluidic chip-Q-TOF with integrated glycan cleavage. ASMS2012).

20. PMM2 gene PCR amplification and sequencing

Primers used for PCR amplification and sequencing of PMM2 gene (OMIM# 601785)

are described in Table 5.

Table 5. Oligonucleotides used for PCR amplification of the PMM2 gene.

Amplicon Primer name Oligonucleotide sequence (5´-3’) (TmºC) Size (bp)

Exon 1 PMM2_Ex1F CCTCCTCTTCCCGACGTGCC (62.5) 254 PMM2_Ex1R CTACCCTGGGTGGTCGATAGC (59.1)

Exon 2 PMM2_Ex2F GACTTATGTACTTGTGTTACCC (50.7)

266 PMM2_Ex2R CTTGGCAACAATATCCTCATAAG (51.6)

Exon 3 PMM2_Ex3F CTGGAGTTTAGCGGTTTTATTG (51.9)

261 PMM2_Ex3R CCTAGAGGCATTCATTGTG (53.4)

Exon 4 PMM2_Ex4F GCTCCTGCTAAATCAAGTAAC (51.2)

218 PMM2_Ex4R CCTATTTGGAGAATGCCCAC (53.4)

Exon 5 PMM2_Ex5F GGAGAAACTCTGTCACCCTT (54.1)

173 PMM2_Ex5R CATAAACCCAGCCATTCACC (53.9)

Exon 6 PMM2_Ex6F CCTACCTTTGTGGCCAGTAG (55.1)

208 PMM2_Ex6R CTCTGGGAAATGGGTATCCAAG (55.1)

Exon 7 PMM2_Ex7F CACCTTTTGCCTTTGTGTGC (55.5)

205 PMM2_Ex7R CCATCAAGCGCAAATGCAAC (55.9)

Exon 8 PMM2_Ex8F CCAGGGTCACATCAGCAATG (56.2) 260 PMM2_Ex8R TGAGCACGTGTGGGAGGACC (61.9)


47

21. Prediction of splice-site, enhancer or silencers, and binding of splicing factors

in the PMM2 gene

Splice scores for the natural and cryptic donor and acceptor splice-sites were

determined using BDGP software (www.fruitfly.org/seq_tools/splice.html). Predictions of

the presence of exonic splice enhancer or silencer sequences were made using ESEfinder

software (http://rulai.cshl.edu/cgi-bin/tools/ESE3) [96], Rescue-ESE [97] and PESX [98]

were used to analyze the binding of several splicing factors.

22. Genotyping of IVS4 +21 G>C and IVS2 -13 G>A PMM2 mutations

Genotyping of two genetic changes identified in PMM2 intronic regions in 2 CDG-like

patients (IVS4 +21 G>C and IVS2 -13 G>A) was done in 61 and 47 healthy subjects by

PCR-allelic specific restriction assay with MnlI enzyme (Fermentas) and by sequencing,

respectively.

23. Platelet purification and proteomic analysis

Blood from one CDG-like patient (P15), one PMM2-CDG patient and four healthy

controls was anticoagulated with 3.8% (w/v) trisodium citrate (9:1) and platelet rich plasma

(PRP) was obtained after centrifugation at 150 g for 15 min. Analysis of platelet

glycoproteins GPIbα, GPIIIa, von Willebrand Factor (vWF), thrombospondin and platelet

endothelial aggregation receptor 1 (PEAR1) was done by western blot using specific home-

made antibodies generated in the laboratory of Dr. K. Freson (University of Leuven,

Belgium), as described before. For proteomic analysis forty mL of blood were collected into

vacutainer ACD tubes (Becton Dickinson, Helsingborg, Sweden) and platelet rich plasma

(PRP) was obtained after centrifugation at 150 g for 15 min.

For proteomic analysis, PRP was transferred to a plastic tube containing PGE1 (1 μM

final concentration), diluted with an equal volume of ACD pH 6.5 (7 mM citric acid, 93 mM

sodium citrate, and 140 mM dextrose) and centrifuged at 150 g for 10 min to remove

contaminant erythrocytes. The supernatant was further centrifuged at 720 g for 10 min and

the platelet pellet was washed twice in JNL buffer (130 mM NaCl, 10 mM trisodium citrate,

9 mM NaHCO3, 6 mM dextrose, 0.9 mM MgCl2, 0.81 mM KH2PO4, and 10 mM Tris, pH

7.4).Glycoproteins from platelet pellets were obtained by lectin affinity chromatography,

using the total glycoprotein spin columns (Qproteome® Total Glycoprotein Kit, Qiagen).

This approach is based on the selective binding capacity to specific glycan moieties of

http://rulai.cshl.edu/cgi-bin/tools/ESE3�


48

specific lectins, namely concanavalin A (ConA) and Wheat germ agglutinin (WGA).

Glycoproteins were then precipitated in thricloroacetic acid in acetone and solubilised in

DIGE buffer (7M urea, 2M thiourea, 4% CHAPS, 30mM Tris, pH8.0).For DIGE analysis,

glycoprotein samples were labelled as previously described [99,100]. Briefly, equal

quantities of platelet proteins from the patients and healthy controls (50 μg for each sample)

were labelled with 400 pmol of Cy3 or Cy5, whereas the pooled internal standard was

labelled with Cy2 (GE Healthcare, Munich, Germany). The samples from any condition

were never labelled with all Cy3 or Cy5 to control any specific dye labelling artefacts that

might occur. The internal standard consisted of a pool of all the samples of the experiments.

Different pairs of Cy3- and Cy5-labeled samples (each containing 50 μg of protein) were

combined and mixed with a 50-μg aliquot of the Cy2-labeled pooled standard. These

mixtures were diluted 1:1 with lyses buffer containing 0.5% IPG buffer (pH 3–10) and 1.2%

Destreak and applied by cup loading on IPG strips (NL pH 3–11, 24 cm). The first

dimension was carried out in an IPGphor system (GE Healthcare) using the following

conditions: 1.5 h at 150 V, 1 h at 500 V in gradient, 2 h at 1,000 V in gradient, 3 h at 8,000

V in gradient and 8 h at 8,000 V. IPG strips were next incubated in equilibration buffer (6 M

urea, 30% glycerol, 2% SDS, and 50 mM Tris–HCl pH 8.8) with 65 mM DTT for 20 min

and in equilibration buffer with 200 mM iodoacetamide and 0.02% bromophenol blue for an

additional 20 min. The second dimensional separations were carried out on 10% SDS-

polyacrylamide gels on the EttanDaltSix system (GE Healthcare).Labelled proteins were

visualized using the Typhoon Trio imager (GE Healthcare). The Cy2, Cy3, and Cy5

components of each gel were individually imaged using excitation/emission wavelengths

specific for Cy2 (488/520 nm), Cy3 (532/580 nm) and Cy5 (633/670 nm). Gel analysis was

performed using DeCyder 2D Differential Analysis Software v6.5 (GE Healthcare).Gels

used to pick spots of interest were visualized with silver staining [101]. The gel pieces were

destained [102] and digested by trypsin (Promega, Madison, WI). Upon concentrating and

desalting tryptic fragments using Millipore C18 ZipTips (Millipore, Bedford, MA, USA),

samples were mixed in a 1:1 ratio with α-cyano-4-hydroxycinnamic acid matrix (saturated

solution in 50% ACN and 2.5% TFA) and spotted onto the target plate. MS/MS analyses

were performed on a 4800 MALDI-TOF/ TOF instrument (Applied Biosystems, Foster City,

CA). The instrument was calibrated with Applied Biosystems Calibration Mixture 1.

Measurements were taken in the positive ion mode between 900 and 3,000 m/z. Sequences

were automatically acquired by scanning first in MS mode and selecting the 15 most intense

ions for MS/MS using an exclusion list of peaks arising from tryptic autodigestion. Data


49

interpretation was carried out using the GPS Explorer software (Version 3.5), and database

searching was carried out using the Mascot program by using MSDB database. MS/MS

searches were conducted with the following settings: MS/MS tolerance for precursor and

fragment ions of 1 Da, carbamidomethylation of cysteine and methionine oxidation as fixed

and variable modification, respectively. Trypsin was selected as enzyme and a maximum of

one missed cleavage was allowed. Using these parameters the probability-based MOWSE

(Molecular Weight Search) scores greater than the given cutoff value for MS/MS

fragmentation data were taken as significant (p< 0.05).

24. Treatment of HepG2 cells with proton and Ca2+ pump inhibitors and a pH

neutralizing agent

Human hepatocellular liver carcinoma cells (HepG2) were treated with 300 nM

bafilomycin A1, a proton-pump chemical inhibitor, 25 mM ammonium persulphate or 50

µM cyclopiazonic acid (CPA), a strong Ca2+ pump inhibitor for 24 hours. Intracellular and

secreted antithrombin was evaluated by immunofluorescence and western blot as described

before.

25. Epidemiologic studies of CDG-like disorders

The epidemiologic studies aiming to determine the prevalence and potential

thrombotic relevance of CDG-like disorders and the PMM2 p.Arg141His mutation was

possible by a multicentre collaboration, and included 1619 patients with deep venous

thrombosis (DVT); 1049 from different Spanish centers (Barcelona, Valencia and Murcia

[87,103,104]) and 570 from Denmark [105]; 791 Spanish patients with Acute Myocardial

Infarction (AMI) [106]. Our study also enrolled 2030 Spanish and 1966 Danish controls.

p.Arg141His genotyping was performed by using a custom SNP Genotyping Assay

(Sigma-Aldrich, Madrid) using Locked Nucleic Acids® (LNA®) which enhance sensitivity

and specificity for the assay. LNA is a novel type of nucleic acid analog containing a 2’-O,

4’-C methylene bridge. PMM2_R141Hwildtype:

[Flc]ACTCA[+A]TG[+C]GT[+T]CT[+T]CTT[TAM]: FLC 5’ fluorophore and TAMRA

quencher; PMM2_R141H mutated: [Rox]ACTCA[+A]TG[+T]GT[+T]CT[+T]CTT[BHQ2]:

Rox fluorophore and BHQ2 quencher. Between brackets the modified LNA nucleotides.

Electrophoretic western blot analyses of plasma samples were carried out to detect

smaller forms of antithrombin and α1-antitrypsin, as described before.


50

Results

51

RESULTS

Results

52

Results

53

Objective 1

1. Anti-FXa activity variability in healthy population

The anti-FXa activity observed in 307 Spanish Caucasian healthy blood donors

had a normal distribution with a mean value of 97±8% (Figure 15). None of the subjects

had antithrombin

deficiency (<70% of the

reference value).

Figure 15. Distribution of

anti-FXa activity in 307

Spanish blood donors.

Subjects with low (L) or high

(H) antithrombin levels

selected for SERPINC1

sequencing are also indicated.

2. Genetic variability of SERPINC1 in healthy subjects with extreme anti-

FXa activity, and validation of potential functional consequences

We selected five subjects with extreme antithrombin levels within the normal

range: three with the lowest levels (L1: 73.5%, L2: 74.5% and L3: 75.1%) and two with

highest levels (H1: 114.4% and H2: 116.1%) (Figure 15). The anti-FXa activity of these

subjects did not significantly change in at least two independent determinations with

different samples. We sequenced 15,375 bp of the SERPINC1 gene, including the

potential promoter, exons, introns and the 3’ region (Figure 16A). Comparison with the

reference sequence revealed 18 polymorphisms and a new g.2085 T>C genetic change

(Figure 16 and Table 6). This substitution, identified in a subject with high antithrombin

levels (H1), was located in the potential promoter region, 228 bp upstream from the

translation initiation codon (Figure 16A). Interestingly, the g.2085 T>C transition could

have potential transcriptional consequences, as it eliminates GATA-1 and c-Myb

binding sites and creates a new site for AP-1 (Figure 16B).

Results

54

Figure 16. Genetic variations identified in the SERPINC1 gene. A) Localization in the SERPINC1

locus of genetic variations identified in this study. The subjects carrying these variations are also

indicated. S: small nucleotide series; L: large nucleotide series: B) Potential transcriptional binding sites

affected by the g.2085 T>C modification located -228 bp from initiation codon. Numbering is based on

genomic DNA sequence, nucleotide +1 corresponds to the A of the ATG initiation codon of the Homo

sapiens antithrombin gene (GenBank accession number NG_012462.1).

The genotyping of this modification in 131 healthy subjects by HRM/Razor probe

method (Figure 14) confirmed that this genetic change is a polymorphism. Actually,

during this study, this change was already described as a SNP in the HapMap

(rs78972925). We identified the g.2085 T>C transition in heterozygous state in seven

out of 131 subjects (5%). However, this polymorphism only showed a trend to be

associated with high anti-FXa activity (99±11% in carriers vs. 96±9% in non-carriers;

p= 0.375) (Table 7).

Results

55

Table 6. Polymorphisms identified in healthy blood donors with low (L1: 73.5%; L2: 74.5% and

L3: 75.1%) and high (H1: 114.4% and H2: 116.1%) antithrombin levels. ins: insertion; I: Intron; E:

exon; WT: Wild-Type allele; ND: not determined; S: small nucleotide series; L: large nucleotide series.

Numbering is based on genomic DNA sequence, nucleotide +1 corresponds to the A of the ATG initiation codon of the Homo sapiens antithrombin gene (GenBank accession number NG_012462.1). * The WT allele contains (ATT)14, while the reported allelic variant contains (ATT)13 in the intron 5 Alu sequence.

We also studied the potential functional effect of the other polymorphisms

identified in this study by correlating genotype with anti-FXa activity in healthy

subjects. This study was restricted to those polymorphisms identified only in individuals

with high or low antithrombin levels. As shown in Table 7, only rs2227589, located in

intron 1, was significantly associated with antithrombin levels. Carriers of the

polymorphic allele had a moderately reduced anti-FXa activity in plasma [65].

Polymorphism Change Location Position H1 H2 L1 L2 L3 g.2085 T>C (rs78972925) T>C 5’ -228 T/C T/T T/T T/T T/T

rs3138521 S>L 5’ -376 S/L S/S S/L ND S/S rs2227589 G>A I 1 182 G/G G/G G/A G/G G/G rs2227590 C>T I 1 361 C/C C/C C/C T/C C/C rs2227592 G>A I 1 686 G/G G/G G/A G/G G/G rs2227596 A>G I 1 1605 A/G A/A A/A ND A/A rs2227603 T>G I 2 3850 T/G T/T T/T ND T/T rs941988 G>A I 2 4505 G/G G/G G/A ND G/G rs941989 G>A I 2 4527 G/A G/G G/A ND G/G rs2227607 G>T I 3 5646 G/T G/G G/G ND G/G rs2227609 G>T I 4 6836 G/T G/G G/G G/G G/G rs5877 A>G E 5 7536 A/G A/A A/G A/G A/A rs5878 A>G E 5 7566 A/G A/A A/G A/G A/A rs1799876 T>C I 5 7927 C/C T/T C/C ND T/T rs5778785 insTTATTA I 5 8073 ins/ins WT WT WT WT rs57195053 (ATT)n* I 5 9231 14/5 14/14 14/11 14/16 14/14 rs2759328 G>A I 5 9693 G/G G/G A/A G/G G/G rs677 G>C I 6 9837 G/G G/G G/G G/C G/G rs2227616 delA I 6 11794 A/- A/A A/A A/A A/A

Results

56

Table 7. Plasma anti-FXa values of blood donors according to the genotype of SERPINC1

polymorphisms identified in selected 5 healthy subjects with extreme antithrombin levels.

Polymorphism Subject Genotyping

method N Frequency Anti-FXa activity (mean±SD) P°

PGA European# Our study

g.2085 T>C (rs78972925)

H1 HRM/Razor probe&sequencing

131 No data T:0.973 WT:96±9 0.375 C:0.027 Pol:99±11

rs2227589 L1 Taqman probe 305 G:0.975 G:0.898 WT:97±7 0.037

A:0.043 A:0.102 Pol:95±8

rs2227590 L2 PCR-ASRA: Fok I

128 C:0.870 C:0.922 WT:95±8 0.793 T:0.130 T:0.078 Pol:94±9

rs2227596 H1 PCR-ASRA: Tsp 45I

127 A:0.795 A:0.886 WT:94±8 0.129 G:0.205 G:0.114 Pol:97±9

rs2227603 H1 PCR-ASRA: Fok I

101 T:0.935 T:0.990 WT:94±8 0.317 G:0.065 G:0.010 Pol:100±19

rs941988 L1 PCR-ASRA: Hinf I

97 G:0.955 G:0.878 WT:97±7 0.513 A:0.045 A:0.122 Pol:95±8

rs2227607 H1 PCR-ASRA: Nla III

138 G:0.972 G:0.895 WT:95±8 0.903 T:0.028 T:0.105 Pol:95±9

rs5778785 H1 HRM&sequencing 106 No data WT:0.896 WT:95±8 0.838

INS:0.104 Pol:95±9

rs2759328 L1 PCR-ASRA: Fau I 149 G:0.917 G:0.850 WT:95±8 0.736 A:0.083 A:0.150 Pol:94±8

rs677 L2 PCR-ASRA: Dde I

158 G:0.842 G:0.930 WT:95±8 0.604 C:0.158 C:0.070 Pol:94±9

rs2227616 H1 PCR-ASRA: Bst EII

183 A:0.935 A:0.880 WT:94±7 0.923 (-):0.065 (-):0.120 Pol:94±8

* N: Number of healthy blood donors studied. # Frequencies from http://www.ncbi.nlm.nih.gov/snp ° p-value obtained from the student-t statistical analysis following a dominant model. WT: Homozygous wild type genotype; Pol: Carriers of the polymorphism (heterozygous + homozygous). SD: standard deviation. rs941988 and rs2227592 were completely linked, according to the Haploview 4.1 The genotypes of all studied polymorphisms were in Hardy-Weinberg equilibrium.

Haplotype analysis revealed 7 frequent haplotypes, although none seemed to be

significantly associated with anti-FXa activity (Figures 17 and 18).

Results

57

Figure 17. Linkage disequilibrium (r2 values) between polymorphisms identified in our study. A

gray scale is used to associate r2 values.

Figure 18. Haplotype analysis. A) Haplotypes identified. WT: Wild-type allele; ins: ATTATT insertion.

B) Correlation between the haplotypes and anti-FXa activity.

Finally, our sequence analysis revealed three new alleles of the rs57195053. This

polymorphism was originally described as a deletion of an ATT triplet from the

sequence of 14 tandem triplets on intron 5. We detected alleles with 5, 11 and 16

triplets. This result together with the description of 11 new polymorphisms affecting

this location (rs72125649, rs71708020, rs71799889, rs71117551, rs71563245,

rs66670532, rs72352633, rs10598596, rs72304917, rs67446132 and rs57195053),

Results

58

supports a single polymorphism with high allelic variability (at least from 5 to 16 ATT

repeats).

3. Genetic analysis in 25 patients with moderate antithrombin deficiency

The clinical features and anti-FXa activity of these patients are shown in Table 8.

In these patients we sequenced the coding region, flanking introns and the potential

promoter region. We only detected three polymorphisms (Table 8). In addition to the

rs3138521, whose potential functional effect was discarded in a previous study [65],

five patients (P1, P7, P18, P20 and P23) carried two linked polymorphisms (rs2227588

and rs61827938) located in the potential promoter region. However, none of these

polymorphisms modified potential binding sites for transcription factors. Moreover, the

functional effect of these polymorphisms was discarded after genotyping in 126 blood

donors (carriers: 93±9; non-carriers: 95±8; p: 0.342).

Table 8. Clinical and genetic features of patients with moderate antithrombin deficiency.

DVT: Deep venous thrombosis. PE: Pulmonary embolism.

N Sex Thrombotic event Age first event

Anti-FXa Alterations

P1 Male DVT 47 72 rs2227588(A/G)* rs61827938 (A/G) P2 Male Recurrent DVT 66 77 No P3 Male DVT 56 88 rs3138521 S/L P4 Male Recurrent DVT + PE 66 86 No P5 Male DVT 59 79 No

P6 Male DVT 35 70 No

P7 Male Recurrent arterial thrombosis 48 78 rs3138521(L/L) rs2227588(A/G)

rs61827938 (A/G) P8 Male Venous thrombosis 42 74 rs3138521 S/L P9 Male Recurrent DVT 65 86 No

P10 Male Recurrent DVT 67 81 No P11 Male Recurrent DVT 68 79 No P12 Male DVT 68 78 No P13 Male Recurrent DVT + PE 72 No

P14 Female Mesenteric venous thrombosis 40 89 rs3138521(S/L)

P15 Male DVT 45 89 No

P16 Male DVT+Arterial thrombosis 47 89 No

P17 Female DVT 60 80 No

P18 Female Cerebral venous thrombosis 27 88 rs61827138 C/T; rs3138521 S/L;

rs2227588 A/G; g.1091C>T P19 Male PE 74 73 No

P20 Male Myocardial acute infraction 36 88 rs61827938 C/T; rs3138521 S/L;

rs2227588 A/G; P21 Male Recurrent DVT 48 81 rs3138521 S/L P22 Male Arterial thrombosis 34 70 No P23 Female Pulmonary thrombosis 24 89 rs2227588 A/G; rs61827938 C/T P24 Male DVT 74 84 No P25 Male Stroke 88 76 rs3138521(S/L)

Results

59

rs3138521: S: small nucleotide series; L: large nucleotide series. * rs61827938 and rs2227588 are in complete disequilibrium linkage.

Interestingly in one patient (P18) we identified a new genetic modification in

heterozygous state not previously described, g.1091 C>T. The presence of this genetic

modification was validated by PCR-allelic specific restriction assay using Rsa I enzyme

24h at 37 ºC. This alteration is located 1,052 bp upstream the translation initiation

codon and close to rs2227588 (only 7 bp away). Further studies are required to clarify

whether these two modifications affect the same allele, although all rs2227588 carriers

identified in this study did not present the g.1091 C>T mutation (Table 8). Finally, in

silico studies predicted that the g.1091 C>T transition abolished a potential upstream

stimulatory factor 1 (USF-1) binding site using both prediction software (TESS and

TFSEARCH), independently of the presence or absence of rs2227588.

4. SERPINC1 analysis in subjects with classical antithrombin deficiency

carrying neither gross gene deletion nor mutation affecting exons or

flanking regions

Patients P26 and P27, with deep venous thrombosis and no familial antithrombin

deficiency, did not present any genetic alteration in SERPINC1 gene including 1,500 bp

from the potential promoter region and 1,000 bp from the 3’ region. Additional clinical

details, biochemical studies and thrombophilia characterization of P27 patient are the

purpose of the third objective.

Family 1 (F1) had severe family history of thrombosis and a type I antithrombin

deficiency. Three sisters with DVT and recurrent events showed 50-56% of functional

antithrombin activity and similar antigen levels. MLPA analysis and sequencing of both

strands of exons and flanking regions were done in two affected members of F1, but no

gross deletion was discovered. Moreover, we did not identify mutations in the rest of

the SERPINC1 gene, including 1,500 bp from the potential promoter region and 1,000

bp from 3’ region.

In contrast, three members of family 2 had a moderate type I antithrombin

deficiency, although no one has developed thrombotic events (Figure 19). The single

relevant genetic change identified in the proband of family 2 was a point mutation

g.2143 C>G, located 170 bp upstream from the translation initiation codon. In silico

prediction suggested that this change could eliminate a potential binding site for the

leader-binding protein-1 (LBP-1) regulatory factor (Figure 19). Then, we genotyped this

Results

60

transversion in 103 healthy blood donors by PCR-allelic specific restriction assay with

Bsa JI, but no one carried it, supporting that this change is not a common

polymorphism. In contrast, the mutation was present in three members of family 2 with

low antithrombin levels (Figure 19B).

In order to further sustain the functional relevance of the g.2143 C>G

transversion, luciferase reporter assays were performed. The plasmid containing the

mutated promoter was associated with 30-50% reduction of luciferase activity in two

different cell lines (HepG2 and PLC-PRF-5) (Figure 19C).

Figure 19. Identification and functional characterization of the g.2143 C>G modification. A)

Electropherogram of the mutation identified in family 2. B) Family 2 pedigree. Proband is pointed by an

arrow. Carriers of the g.2143 C>G modification are represented with the semi filled symbol. ND: not

determined. Plasma anti-FXa activities (%) are also shown. C) Luciferase reporter assay of SERPINC1

promoter activity with the wild-type and mutated g.2143G allele in HepG2 and PLC-PRF-5 cell lines.

Results were obtained from 5 independent experiments.

Results

61

Objective 2

The results of the previous objective have provided quite clear evidences about

the low genetic variability in SERPINC1 gene and the minor functional consequences to

antithrombin levels. These findings encouraged the search for other potential

modulating genes of plasma antithrombin levels by using a GWAS approach, aiming to

find associations between genotypes and plasma anti-FXa activity.

1. GWAS analysis. Association between genotype and plasma anti-FXa

activity in the GAIT study

The plasma anti-FXa activity in the GAIT study had a normal distribution, with a

medium value of 109.05% of the reference plasma and 154% and 78% as extreme

values. Results of the GWAS are summarized in figure 20. No SNP was found

associated with plasma anti-FXa activity at a genome-wide significance level (figure

20). For validation analysis we selected the 10 SNPs with the strongest association (p<

4x10E-05). Interestingly, 2 additional polymorphisms affecting LARGE, one of the

genes identified, also showed significant association with anti-FXa activity, and were

also selected to be validated (rs762057 and rs240082). Two of the LARGE SNPs

(rs713703 and rs762057) displayed high linkage disequilibrium (D2= 0.81). Table 9

displays the list of these polymorphisms, showing the p-value for the association with

anti-FXa activity, the chromosomal location, and the gene potentially affected.

Figure 20: Manhattan plot GWAS with antithrombin phenotype. The threshold of significance to

select candidate SNPs for validation is also shown.

Results

62

Table 9. Single nucleotide polymorphisms (SNPs) associated with anti-FXa activity in the GWAS of

the GAIT study and that were selected for validation studies.

SNPS PVAL BSNP V-EXPL CHR LOC GENE

rs10880942 0.00000044 -0.734718 0.081805 12 44905397 SLC38A1

rs2356895 0.00000128 -0.454476 0.088341 14 50898162 LOC283553

rs1860867 0.0000141 0.741308 0.050958 7 51762794 COBL

rs9896932 0.0000191 -1.361577 0.028271 17 77848326 CD7

rs1411771 0.0000206 0.361303 0.063374 1 230241398 DISC1

rs2152192 0.0000277 0.765150 0.061933 6 130510819 SAMD3

rs13193455 0.0000295 0.362923 0.029200 6 170354404 LOC154449

rs713703 0.0000313 0.330586 0.058187 22 32286080 LARGE

rs11681944 0.0000328 -0.338764 0.062070 2 68176083 C1D

rs6768189 0.0000357 0.429999 0.051721 3 54446050 CACNA2D3

rs762057 0.0000944 0.310900 0.052599 22 32280513 LARGE

rs240082 0.0010970 0.707098 0.039564 22 32428417 LARGE

2. Validation study

The 10 SNPs that showed stronger statistical association with anti-FXa in the

GWAS, as well as the 2 additional LARGE SNPs were genotyped in 307 blood donors

from another Spanish region used as a validation cohort. Only rs762057 maintained a

significant association with anti-FXa levels in the validation cohort (p= 0.047) (Table

10). Multivariate analysis including age, gender and rs2227589, a SNP in SERPINC1

gene previously reported to be associated with plasma anti-FXa activity [65], increased

the significance between rs762057 and anti-FXa activity (p= 0.020) (Table 10).

Results

63

Table 10. Genotype-phenotype analysis in the validation study.

SNP Gene Genotype Anti-FXa Crude p Adjusted p 1. rs10880942 SLC38A1 T/T 96.3±7.4 T/C+C/C 97.1±8.1 0.444 0.414 2. rs2356895 LOC283553 T/T 96.2±7.4 T/C+C/C 96.8±7.8 0.510 0.240 3. rs1860867 COBL G/G 96.7±7.4 G/A+A/A 95.1±8.4 0.178 0.240 4. rs9896932 CD7 A/A 96.3±7.5 A/G+G/G 96.7±8.3 0.800 0.870 5. rs1411771 DISC1 T/T 95.7±7.9 T/C+C/C 97.0±7.2 0.122 0.078 6. rs2152192 SAMD3 G/G 96.4±7.6 G/A+A/A 96.2±7.4 0.839 0.998 7. rs13193455 LOC154449 G/G 96.2±7.5 G/A+A/A 96.6±7.6 0.651 0.705 8. rs762057 LARGE G/G 95.3±6.9 G/A+A/A 97.1±7.7 0.047 0.020 9. rs11681944 C1D G/G 96.3±8.0 G/A+A/A 96.6±7.1 0.730 0.655 10. rs6768189 CACNA2D3 G/G 96.8±7.8 G/A+A/A 95.6±6.9 0.158 0.198 11. rs713703 LARGE T/T 95.2±7.2 T/C + C/C 96.8±7.5 0.082 0.087 12. rs240082 LARGE A/A 96.4±7.7 A/G + G/G 96.0±6.9 0.800 0.733

Finally, LARGE haplotype analysis in the validation cohort revealed 5 frequent

haplotypes, one of them (H2) significantly associated with anti-FXa activity (p= 0.030)

(Table 11).

Table 11: LARGE haplotypes identified in the validation study and their correlation with anti-FXa

activity.

Haplotype rs762057 rs713703 rs240082 Frequency Difference (95% CI) P

1 G T A 0.4523 0.00 ---

2 A C A 0.403 1.4 (0.12-2.68) 0.033

3 G C A 0.0666 2.24 (-0.34-4.81) 0.09

4 A T A 0.0399 1.8 (-1.36-4.95) 0.27

5 G T G 0.0258 1.18 (-3.17-5.52) 0.6

rare * * * 0.0124 -1.42 (-8.16-5.31) 0.68

Results

64

Consistently in both cohorts, the polymorphic A allele associated with a moderate

higher plasma anti-FXa activity (2-4%) (Figure 21).

rs762057 (LARGE)

GAIT Study Validation Cohort

90

100

110

120G/GG/A+A/A*

*

Ant

i-FX

a ac

tivity

(%)

Figure 21: Plasma anti-FXa activity according to the rs762057 genotype in the GAIT and validation

studies. *p<0.05

3. Thrombotic relevance of the LARGE rs762057 polymorphism

Although the A allele only mildly correlated with higher antithrombin levels, the

key relevance of this anticoagulant encouraged to evaluate whether or not the rs762057

polymorphism affected the risk of venous thrombosis in our case-control study.

Genotyping was successful in 849 patients and 862 controls. The allele or genotype

frequencies were similar in cases than in controls (Figure 22).

Results

65

rs762057 (A>G)

AA GA GG G A0

20

40

60ControlsPatients

Freq

uenc

y (%

)

Figure 22. Genotype and allelic frequencies of rs762057 in the case/control study including 849

patients with venous thrombosis and 862 controls.

Actually, both the univariate and multivariate analysis including age, gender and

genetic prothrombotic risk factors (FV Leiden, prothrombin G20210, antithrombin

Cambridge II A384S and ZPI Arg67Stop) revealed that this polymorphism did not

significantly modify the risk of venous thrombosis (p= 0.92 and p= 0.94, respectively).

4. Functional studies

In order to verify the potential role of LARGE as a modulating gene of

antithrombin identified by genetic association studies, further functional studies were

performed.

Firstly, as the human LARGE gene is ubiquitously expressed, with highest levels

in heart, brain and skeletal muscle [107], we evaluated the expression of this gene in

mononuclear cells from 10 healthy subjects. Interestingly, the expression of this gene in

mononuclear cells showed a significant and positive correlation (r2= 0.8) with plasma

anti-FXa activity (Figure 23A).

Since LARGE codes an enzyme involved in post-translational glycosylation, and

glycosylation of antithrombin plays a relevant role in the function of this serpin,

particularly in the heparin affinity [24,77,108], our first hypothesis considered that

differential expression or function of LARGE could result in distinct glycosylation of

antithrombin. In order to verify this hypothesis, proteomic and glycomic studies were

done with the main plasma antithrombin glycoform (α-antithrombin) purified from

subjects with the highest and lowest LARGE expression. However, their molecular

Results

66

masses were very similar, and glycomic studies showed fluctuations but not significant

differences on the level or type of glucidic components (Figure 24). These results

suggested that the association with anti-FXa activity might be explained by a

quantitative defect rather than by qualitative differences caused by the differential

LARGE expression. Actually, plasma antigen levels of antithrombin also significantly

correlated with LARGE expression in these healthy subjects (Figure 23B).

0.00 0.01 0.02 0.0380

90

100

110r2= 0.8

LARGE expression

Ant

i-FX

a ac

tivity

(%)

0.00 0.01 0.02 0.0380

90

100

110r2=0.4

LARGE expression

Ant

igen

leve

ls o

f ant

ithr

ombi

n (%

)

A) B)

Figure 23: Correlation between mononuclear LARGE expression and plasma anti-FXa activity (A) and

antigen levels (B) in 10 healthy subjects. LARGE expression was evaluated by specific transcription probe.

Values are expressed as % of the reference value obtained in a pool of 100 healthy subjects.

Results

67

Results

68

Figure 24. Glycomic and proteomic analysis of healthy subjects with the highest (blue) and lowest

(red) LARGE expression. The main glycoform α (black), and the minor β glycoform (green), both

purified from a pool of 100 healthy blood donors are also shown. The β glycoform has 3 N-glycans since

it lacks N-glycosylation at N-135. A) MALDI TOF mass spectrometric analysis of: 1) Intact

glycoproteins; 2) 2AB-labeled N-glycans. B) HPLC data. 1) Distribution of the glycan structures of

antithrombin specimens. Values are represented as % of total glycan pool. Between brackets are the

absolute fluorescence units. 2) HILIC HPLC profiles of antithrombin specimens.

Unfortunately, these results can be questioned because of the weak expression of

LARGE in mononuclear cells and the moderate differences found in healthy subjects

with the highest and lowest LARGE expression (6.2-fold: 0.028 and 0.0045 units

relatives to the expression of the constitutive gene, respectively).

To strongly sustain the relevance of LARGE on antithrombin levels, we carried

out silencing experiments of LARGE in HepG2 and HEK-EBNA cell lines. According

to our data, 60 hours after transfection with LARGE silencers, the expression of this

gene was silenced around 30% in both cell lines (Figure 25). At this time, secretion of

antithrombin to the conditioned medium in both cell lines treated with LARGE silencers

was significantly reduced in silenced cells, 4-fold by western blot (Figure 26A) and 10-

fold by ELISA (0.01 ± 0.01 mg/ml compared to 0.15 ± 0.20 mg/ml of control cells).

The reduction was more significant in HEK-EBNA cells (Figure 26A). However,

according to electrophoretic data, secreted antithrombin from silenced cells shows

similar size to that of control cells (Figure 26A Interestingly, anti-FXa activity in the

conditioned medium of LARGE silenced cells was 59 ± 30% and 11 ± 12% of that

found in control cells transfected with the scramble siRNA or without siRNA in HepG2

and HEK-EBNA respectively. The reduction of antithrombin secretion paralleled with a

moderate intracellular retention of this serpin according to the immunofluorescence and

western blot results (Figure 26B).

Results

69

HEK-EBNA HepG2

60

80

100ControlsiRNA

mR

NA

LARG

E(%

)

Figure 25: LARGE gene expression in HEK-EBNA and HepG2 cell lines 60 hours after transfection

with LARGE siRNAs. Values are expressed as percentages respect to control cells (100%) (N= 3).

Moreover, aiming to determine the mechanisms underlying the modulation of

antithrombin by LARGE, we measured SERPINC1 expression in these cells. Silencing

of LARGE did not significantly modify SERPINC1 mRNA levels (Figure 26C).

In HepG2 cells, we also studied the effect of LARGE silencing on other N-

glycoproteins: prothrombin, transferrin and other hepatic serpin: α1-antitrypsin. As

shown in figure 26A, silencing of LARGE also reduced the secretion of all other

proteins evaluated, although antithrombin seemed to be the most affected.

Results

70

Figure 26. LARGE gene silencing in HepG2 and HEK-EBNA cell lines. A) Consequences on proteins

secreted to the conditioned medium evaluated by immunoblotting. B) Effect on intracellular antithrombin

from HepG2 cells analyzed by immunofluorescence and immunoblotting. C) Effect on the levels of

SERPINC1 expression in HEK-EBNA and HepG2 cell lines. Immunoblots and immunofluorescence

figures are representative of at least 3 independent experiments.

Results

71

Objective 3

The analysis of our large cohort of 89 patients with thrombosis and severe

deficiency of antithrombin has identified the mutations or deletions responsible for

antithrombin deficiency in 95% of these cases. However, the remaining 5% of patients

with antithrombin deficiency and no mutation in the SERPINC1 gene is an excellent

group to identify new mechanisms responsible for the deficiency of this important

anticoagulant.

1. Diagnosis of a congenital disorder of glycosylation (PMM2-CDG) in a

patient with antithrombin deficiency and recurrent venous thrombosis

Our first study focused on three patients with thrombosis and severe antithrombin

deficiency but no genetic defect on the SERPINC1 gene. Patient P26 and F1 had type I

antithrombin deficiency. Western blot analysis only confirmed the reduced levels of

antithrombin in plasma, but did not reveal abnormal antithrombin forms.

Patient P27 is an 18 year old Portuguese male with recurrent left-leg venous

thrombosis. The first event happened at 11 years old without any known concomitant

risk factor, including fever or immobilization. Oral anticoagulation with warfarin was

established for 5 months. The patient suffered a second spontaneous event in the same

localization at 12 years old. Oral anticoagulation with warfarin was re-established.

Unexpectedly, a new thrombotic event took place 6 months later under stable oral

anticoagulant therapy with INR 2-3. A fourth thrombotic event was diagnosed at 17

years old, again under oral anticoagulant therapy with INR 2-3. Now, the patient is

being treated with higher warfarin dosis (INR 2.5-3.5).

No familial history of thrombosis was reported. Thrombophilic analysis revealed

no prothrombotic polymorphisms (Factor V Leiden or prothrombin G20210A). No

antiphospholipid antibodies were detected. Protein C and protein S had values within

the normal range. The single thrombophilic defect identified was an antithrombin

deficiency (67% plasma anti-FXa activity versus 100% in a pool of plasma from 100

healthy blood donors). Antigen levels of antithrombin in plasma were also low (60%),

suggesting a classical heterozygous type I antithrombin deficiency. However, the

sequencing of the potential promoter, as well as the coding and flanking regions of the

SERPINC1 gene [109] revealed no significant mutation. Large gene deletions were

excluded by MLPA studies. Interestingly, electrophoretic analysis of plasma

Results

72

antithrombin performed on denaturing and native conditions revealed the presence of an

abnormal antithrombin (Figure 27A). The reduced size of this abnormal band, observed

in SDS gels under reducing conditions, and the absence of mutations in the coding

region encouraged the search for a post-translational modification. We evaluated an

impaired N-glycosylation, the only post-translational modification described in

antithrombin that also affects the function and clearance of this anticoagulant serpin

[108]. We treated plasma from the patient and a control with N-glycosidase F, as

described before. As shown in figure 27B, N-glycan removal uniformed the size of

plasma antithrombins in SDS gels sustaining an N-glycosylation defect for about 20%

of antithrombin molecules in the plasma of the proband.

Results

73

Figure 27. Identification of PMM2-CDG in the proband (P27). A) Native and SDS-PAGE under

reducing conditions showing the normal (arrow) and abnormal (dashed arrow) antithrombin present in

plasma of the proband, a patient with congenital type I antithrombin deficiency (AT def) (SERPINC1

p.Arg117Stop) and a control subject. B) SDS-PAGE and western blot of antithrombin from plasma

untreated (-) or treated (+) with N-glycosidase F. C) SDS-PAGE pattern of α1-antitrypsin (α1-AT) and

antithrombin (AT) in the proband, his mother, a control and a PMM2-CDG patient (PMM2 p.Phe119Leu

and p.Arg141His). Arrows mark the abnormal glycoforms of these molecules. D) HPLC profile of

transferrin glycoforms of the proband, his mother and a PMM2-CDG patient (PMM2 p.Phe119Leu and

p.Arg141His).

This abnormal antithrombin was not the beta-glycoform as it has different

electrophoretic mobility in SDS-PAGE (Figure 28) and among other evidences, this

defect was not specific of antithrombin as smaller forms of α1-antitrypsin were also

detected in plasma of this patient (Figure 27C).

Figure 28. Electrophoretic pattern of plasma antithrombin from a control and the proband (P27)

compared to that of a mixture of alpha and beta antithrombin (AT) purified from plasma of

healthy subjects.

All these biochemical and clinical data focus the search to congenital disorders of

glycosylation (CDG), a broad group of autosomal recessive disorders [110]. Actually,

antithrombin and α1-antitrypsin electrophoretic profiles were very similar to that

displayed by a patient with the most common CDG: PMM2-CDG, previously called

CDG-Ia (Figure 27C). PMM2-CDG patients usually are compound heterozygous for

mutations in PMM2, the gene encoding phosphomannomutase 2, a cytosolic enzyme

that catalyzes the conversion of mannose 6-phoshate to mannose 1-phosphate, a

substrate for GDP-mannose synthesis, required in N-glycosylation [110]. HPLC

analysis of transferrin glycoforms pointed to a CDG-Ia diagnosis by the presence of

asialotransferrin and a significant increase of disialotransferrin (Figure 27D). PMM2

sequencing in the proband identified two mutations: one of maternal origin in exon 5,

α

α/β Proband Control

β

Abnormal AT

Results

74

and the other in exon 8, responsible for the p.Arg141His and p.Cys241Ser amino acid

substitutions, respectively (Figure 29). The same combination has been reported

previously in a PMM2-CDG patient with mild clinical phenotype [111]. In addition, the

proband had strabismus, a thyroid nodule, kyphosis and moderate mental retardation

without ataxia. The digestive, hepatic, renal and cardiovascular systems were normal.

Figure 29. Electropherograms of mutations identified in the PMM2 gene of the proband (P27).

2. Identification of a CDG-like disorder

The diagnosis of a PMM2-CDG disorder in a patient with antithrombin deficiency

but without mutation in the SERPINC1 gene encouraged the search for this o similar

defect among the 25 patients with venous thrombosis and a moderate reduction of anti-

FXa activity (70-90%, Table 8) but no mutations in SERPINC1 gene. Thus, we studied

their plasma antithrombin by electrophoretic analysis, including native and SDS gels

under reducing and no reducing conditions.

Interestingly, four patients (P2, P15, P16, and P22) showed abnormal plasma

antithrombin pattern in SDS-PAGE under reducing conditions (Figure 30A). They

presented a smaller band not present in controls. The amount of this abnormal band

never reached more than 15% of total antithrombin although varied in these patients,

being P15 the patient with higher levels (Figure 30A). Purification by FPLC of this

smaller antithrombin from P15, followed by mass spectrometry quantification revealed

an abnormal antithrombin of 56.4 kDa. Treatment of plasma of this patient and a control

with N-glycosidase F uniformed abnormal bands, suggesting again an N-glycosylation

defect. Further analysis of plasma antithrombin from this patient by treatment with

neuraminidase and galactosidase revealed incomplete sialic acid and galactose content

Results

75

(Figure 30B). Then, we compared the electrophoretic pattern of these patients with a

PMM2-CDG patient. As illustrated in figure 30C for P15, the pattern of antithrombin,

α1-antitrypsin and transferrin glycoproteins was really similar. Further evidences

strengthen the similarities of these patients with PMM2-CDG patients. Thus, HPLC and

Q-TOF also confirmed the presence of abnormally glycosylated proteins in these

patients. Transferrin glycoform analyses by HPLC revealed the presence of

asialotransferrin and a significant increase of disialotransferrin in these patients (Figure

31). Moreover, Q-TOF analysis showed increase levels of aglycan and monoglycan

structures of transferrin (Figure 32). Importantly, the electrophoretic and HPLC patterns

of parents of PMM2-CDG patients with a single PMM2 mutation displayed was similar

than that of controls (Figure 27).

Figure 30. Electrophoresis analysis in SDS-PAGE under reducing conditions of patients with a

CDG-like disorder (P15, P16, P2 and P22). A) Electrophoretic pattern of plasma antithrombin. B)

Treatment of plasma antithrombin with N-glycosydase F (N-Glyco F), neuraminidase (Neu) and

galactosidase (Galac). C) Electrophoretic pattern of antithrombin, α1antitrypsin and transferrin of P15

comparing with a control and a PMM2-CDG patient. C: Control.

Results

76

Figure 31. HPLC profiles of plasma transferrin glycoforms of CDG-like patients (P15, P16 and P2),

a control and a PMM2-CDG patient.

Results

77

Control

P15

P16

PMM2-CDG

Results

78

Figure 32. Q-TOF mass spectrometry of transferrin glycoprotein of CDG-like patients (P2, P15 and

P16), a PMM2-CDG patient and a control. Peaks correspond to transferrin with 0, 1, or 2 N-glycans.

Data are expressed as ratio (aglyco = 0 glycans; monoglyco = 1 glycan; diglyco = 2 glycans).

Finally, PMM2 sequencing identified a different heterozygous mutation in three

out of four CDG-like patients: P15, P16 and P2 (c.422 G>A -p.Arg141His-, IVS4 +21

G>C and IVS2 -13 G>A, respectively) (Figure 33).

Figure 33. Electropherograms of PMM2 mutations found in CDG-like patients (P15, P16 and P2).

The p.Arg141His mutation has a strong functional effect and is the most frequent

genetic defect identified in PMM2-CDG patients, particularly in the north of Europe

[112]. The remaining two mutations, IVS4 +21 G>C and IVS2 -13 G>A, could have

significant effects on the splicing process according to in silico modelling as both

generate new branch sites with 1.5 and 2.2 scores, respectively (threshold= 0).

Moreover, ESE Finder software identified two potential serine/arginine-rich proteins

(SR proteins) which are involved in regulating and selecting splice sites (SRSF5 and

SRSF1) able to bind to the mutated sequence of IVS4+21 G>C with 3.08 and 3.07

scores, respectively.

We discarded these PMM2 mutations as common polymorphisms in the Spanish

population since the p.Arg141His substitution was found in 5 heterozygous subjects out

of 1,570 healthy controls (0.3%, allele frequency 0.0016), and the IVS4 +21 G>C and

IVS2 -13 G>A mutations were not identified in 61 and 47 healthy subjects, respectively.

P2

Results

79

However, these intronic mutations have never been described in PMM2-CDG patients

[112].

Hence, both a similar laboratory phenotype and a common clinical feature

(thrombosis) of these patients with a single PMM2 mutation suggest the existence of a

new disorder that we have called CDG-like.

3. Follow-up of patient P15

Five years follow-up of patient P15 allowed us to study the evolution of abnormal

antithrombin, α1-antitrypsin and transferrin glycoforms. As shown in figure 34, the

amount of abnormal α1-antitrypsin glycoforms is highly variable, being hardly

detectable at certain time points. As expected, there is a positive correlation between the

amounts of these abnormal α1-antitrypsin glycoforms with the peak of asialotransferrin.

Interestingly, we also detected an inverse correlation of these parameters with plasma

antithrombin activity (Figure 34).

Figure 34. Evolution of α1-antitrypsin glycoforms, plasma antithrombin (AT) activity (% of anti-

FXa compared with a pool of 100 healthy controls) and asialo-transferrin (area of the

asialotransferrin peak in HPLC) levels in patient P15 with CDG-like during five years of follow-up.

The date of each sample (month-year) is indicated. C: healthy control.

4. Proteomic analysis of platelet glycoproteins in PMM2-CDG and CDG-

like patients

All abnormal glycosylated proteins identified in CDG-like and PMM2-CDG

patients were of hepatic origin. We wanted to study whether these disorders could also

affect N-glycoproteins from other cells or tissues. Most studies were done with platelet

proteins due to the key haemostatic role of platelets. Western blot analysis did not

Results

80

reveal abnormal glycoforms of GPIbα, GPIIIa, von Willebrand factor (vWF),

thrombospondin or PEAR1 neither in PMM2-CDG nor CDG-like patients (Figure 35).

We were not able to detect smaller glycoforms of many other proteins from different

origin: immunoglobulin, plasminogen activator inhibitor 1 (PAI-1) or tissue factor

pathway inhibitor (TFPI) (Data not shown).

Figure 35. Western blot of platelet glycoproteins in a PMM2-CDG patient, a CDG-like patient

(P15) and two controls.

We also performed a proteomic study of platelet glycoproteins. The 2D pattern of

a PMM2-CDG patient was similar to that of a CDG-like patient (P15), but they were

different to that of controls (Figure 36). Interestingly, MALDI-TOF analysis of tryptic

peptides from unique spots of PMM2-CDG and CDG-like patients, which were

compatible with a glycosylation defect due to the smaller size and increased pI; revealed

the presence of hepatic proteins: α1-antitrypsin, transferrin or fibrinogen (Figure 36).

However, this study did not identify any abnormal platelet glycoprotein in PMM2-CDG

or CDG-like patients.

Results

81

Figure 36. 2D gel of platelet glycoproteins from a control, a PMM2-CDG patient and a CDG-like

patient (P15). Arrows show spots further analysed by mass spectrometry. Peptides identified correspond

to hepatic proteins: 1 transferrin 2: β-fibrinogen, 3: α1-antitrypsin.

5. Search for additional factors involved in CDG-like disorders

The identification of a single PMM2 defect in 3 patients with CDG-like disorders

together with the fact that these patients do only share a clinical feature with PMM2-

CDG patients (thrombosis) strongly suggest the presence of an additional factor that

exacerbate the effects of the single PMM2 mutation. P15 also had Darier disease which

was inherited by his mother. We originally hypothesized that this combination might

explain the CDG-like disorder. We speculated that the mutation of the Ca2+ pump that

features the Darier disease [113] might finally affect the pH of Golgi, leading to the

mislocalization of galactosyl-transferases, as described for inhibitors of the proton pump

or pH neutralizing agents [114]. In order to test this hypothesis we treated HepG2 cells

with bafilomycin A1, a proton-pump chemical inhibitor. As indicated in figure 37, this

inhibitor caused a significant reduction in the levels of antithrombin secreted to the

medium, potentially by its toxic effects on this cells, but the antithrombin secreted

showed an abnormal glycosylation. Similarly, ammonium persulphate, an agent that

neutralizes the pH inside Golgi, caused formation of endosomes (Figure 38), as

previously described [114], and reduced glycosylation of secreted antithrombin (Figure

37). However, treatment of HepG2 cells with cyclopiazonic acid (CPA), a strong Ca2+

pump inhibitor, had a toxic effect that reduced the levels of secreted antithrombin, but

had negligible effects on antithrombin glycosylation (Figure 37).

1

23

1

23

Control PMM2-CDG P15

Results

82

Figure 37. Electrophoretic analysis of antithrombin secreted to the conditioned medium of HepG2

treated with bafilomycin A1 (Baf), anmonium persulphate (NH4Cl) and cyclopiazonic acid (CPA);

α/β a mixture of α and β glycoforms of antithrombin purified from plasma of healthy subjects.

Figure 38. Immunoflourescence of intracellular antithrombin from HepG2 cells treated with

anmonium persulphate (NH4Cl).

Moreover, no further CDG-like patient reported Darier disease. In contrast, and

interestingly, all four CDG-like patients identified in this study reported high alcohol

consumption.

6. Epidemiologic analysis evaluating potential CDG-like disorders in patients

with thrombosis

We firstly evaluated the prevalence of CDG-like disorders in two different

thrombotic diseases: DVT and AMI evaluating the electrophoretic pattern of α1-

antitrypsin in plasma samples from 571 Spanish patients with DVT, 262 Spanish

patients with AMI and 393 Spanish controls. Patients and controls with an abnormal

Results

83

pattern compatible with a CDG-like disorder were verified by electrophoretic analysis

of plasma antithrombin. A mild CDG-like pattern was identified in 4 DVT patients, 2

AMI patients and only one control. Table 12 indicates the clinical data and results of

PMM2 sequencing in these subjects. Interestingly, the control and one DVT patient

carried the p.Arg141His mutation and one patient carried a new mutation in intron 2 of

PMM2 (IVS2-14insT) in compound heterozygosity with Glu197Ala.

Table 12. Clinical and genetic features of patients and controls with a CDG-like pattern. DVT:

patients with deep venous thrombosis; AMI: patients with acute myocardial infarction; WT: Wild

Type. The age of the first event is between brackets.

Patient Gender Age Recurrence Risk factors

Familial history PMM2

DVT1 Male 42(40) No No Yes WT

DVT2 Female 47(38) No FVL No E197A&IVS2-14InsT*

DVT3 Male 35(34) Yes No ND WT

DVT4 Female 36(36) No No No R141H

AMI1 Male 42(42) No No Yes rs11074924 (T/T) AMI2 Male 48(48) No No No rs11074924 (T/T)

Control Male 36 No No No R141H * New mutation in intron 2 of PMM2

Then, we screened the prevalence of p.Arg141His, the most frequent PMM2

alteration identified in PMM2-CDG patients, also identified in CDG-like patients. The

genotyping was done in two case-control studies from different populations with

different allele frequency for this mutation: Denmark and Spain. Figure 39 shows the

results of this study. The p.Arg141His mutation had higher prevalence in patients with

thrombosis than controls but differences did not reach statistical significance.

Results

84

Danish population Spanish population0

1

2

3

ControlsPatients

p=0.288OR:1.41; 95% CI (0.71-2.71)

p=0.262OR:1.85; 95% CI (0.57-6.35)

(%) A

rg14

1His

pre

vale

nce

Figure 39. Prevalence of PMM2 p.Arg141His mutation in patients and controls.

Discussion

85

DISCUSSION

Discussion

86

Discussion

87

The great importance of antithrombin in the physiopathology of the haemostatic

system is due, in some extent, to the so efficient anticoagulant mechanism of this serpin

and the wide variety of procoagulant proteases that is able to inhibit [115]. Furthermore,

since 1965 when O Egeberg described the first family with deficiency associated with a

high incidence of venous thrombosis [2] more than 296 different mutations have been

identified in patients with venous thrombosis and antithrombin deficiency. Although the

clinical features of carriers of congenital antithrombin deficiency may be quite

heterogeneous, the thrombotic risk associated with antithrombin deficiency is in general

high, particularly in patients with type I deficiency whose risk is about 20-50 fold

respect healthy people [57]. Fortunately, the prevalence of antithrombin type I

deficiency is very low, even among patients with venous thrombosis (1 %) [116]. In

contrast, subjects carrying the antithrombin Cambridge II mutation (p.Ala384Ser),

relatively frequent (0.2% in our population), have a moderate risk of thrombosis,

because it only mildly impairs the anti-FIIa activity of antithrombin [76]. In agreement

with these data, moderate reduction in antithrombin levels has recently been suggested

to mildly increase the thrombotic risk [70]. All these data strongly encourage

investigating all potential mechanisms that might impair antithrombin levels or

function. This search has rendered excellent results in patients with congenital

antithrombin deficiency, as the causative mutation is identified in a high percentage of

cases (up to 90% according to recent large studies [74], and 95% in our large cohort of

patients with antithrombin deficiency). However, these studies are restricted to the

SERPINC1 gene, the gene encoding antithrombin, and all genetic defects identified are

gross gene deletions or mutations affecting exons or flanking regions. The practically

absence of genetic modifications associated with reduced levels of this key

anticoagulant through regulatory mechanisms is surprising, probably reflecting that

mutations affecting these regions are very rare or have small functional consequences.

Moreover, few regions with potential regulatory role have been described in the

SERPINC1 gene. So far, only one report from our group has reported a very moderate

effect associated to a common polymorphism in intron 1 [65], which also explains its

negligible thrombotic consequences [117-119].

This study aimed to identify new genes and genetic modifications involved in the

regulation or modulation of antithrombin levels or function, and to characterize the

mechanism(s) involved. To achieve this objective, we followed three different

approaches:

Discussion

88

1. To identify new regulatory regions in the SERPINC1 gene, we sequenced a

15,000 bp of this gene in three groups of samples with different antithrombin

levels.

2. To identify antithrombin modulating genes, a Genome Wide Association

Study (GWAS) and further experimental validations were done.

3. To identify new mechanisms leading to antithrombin deficiency, we studied

the biochemical features of plasma antithrombin from patients with

antithrombin deficiency and no mutations in SERPINC1 gene.

The result of this study reveals a fascinating complex heterogeneity of new

mechanisms affecting the levels of antithrombin with potential pathological

consequences.

1) We analyzed 1,500 bp of the promoter region of the SERPINC1 gene in 29

subjects with antithrombin deficiency without genetic modifications in the coding or

splicing regions and no gross gene deletions. This study identified the first mutation in

the promoter region associated with antithrombin deficiency. Certainly, the g.2143 C>G

change identified in family 2 does not abolish the transcription of the mutated allele, as

demonstrated by the antithrombin levels observed in carriers, but significantly impairs

it. This functional effect was further confirmed by reporter assays in different cell lines.

Moreover, our results verify the regulatory relevance of the region of at least 200 bp

upstream the ATG start codon. The g.2143 C>G modification is located in one of the

four protected regions identified in two independent reports [61,120]. Thus, it is located

at the end of region C proposed by Fernandez-Rachubinski and coworkers [61], and in

the middle of region II according to the nomenclature of Tremp and co-workers [120]

(Figure 40). Interestingly, it has been suggested that the hepatic nuclear factor 4 (HNF4)

interacts with this region [61] (Figure 40).

Discussion

89

Figure 40. Localization of g.2143 C>G in potential promoter regulatory regions of the SERPINC1

gene identified by DNase protection assay. Sequences highlighted in grey (A-C) were described by

Fernandez-Rachubinski and coworkers [61] while regions underlined (I-IV) were described by Tremp and

coworkers [61,120]. The interacting transcription factors identified are indicated below each sequence

Moreover, preliminary results in one additional patient with moderate

antithrombin deficiency have identified one new mutation g.1091 C>T, with a potential

regulatory effect in the most distal promoter. Further familial and reporter studies are

required to verify the regulatory relevance of this genetic modification as well as to

identify the mechanism(s) involved.

Therefore, our results support that this promoter region should be included in the

molecular analysis of subjects with antithrombin deficiency, particularly in those

subjects with type I deficiency and without mutation in the coding region of SERPINC1.

However, we did not detect mutations or any functional polymorphism in the

promoter of 24 patients with moderate antithrombin deficiency or two patients with

severe antithrombin deficiency. These data strongly suggest the presence of additional

regulatory elements located up or down stream the region studied. Actually, the

identification of a family (F1) and one patient (P26) who belongs to the same

geographic area (Galicia) than F1, with type I antithrombin deficiency and severe

thrombosis who had neither gross deletion nor functional mutations in the 15,925 bp of

Discussion

90

the SERPINC1 gene, encourage further search for remote regulatory regions. The recent

ENCODE project may assist this search [121].

Regulatory regions affecting the SERPINC1 gene might also be studied in the

general population. Actually, the anti-FXa activity, the method widely used in

thrombophilic test to detect antithrombin deficiency, shows a great variability in the 307

healthy donors with normal distribution (70-130%) [122]. Factors such as sex, body

mass index, oral contraceptives or race have been associated with antithrombin

variability [90], but genetic factors must also play a relevant role according to the high

heritability of this trait (h= 0.486) [64]. The sequence of the SERPINC1 locus in five

subjects with extreme antithrombin levels within the normal range (75% and 115%)

selected from 307 healthy donors aimed to detect genetic variations in the SERPINC1

gene involved in such a great interindividual variability. We verified the minor

functional effect of the rs2227589 polymorphism located in intron 1, and we found a

polymorphism in the promoter region that deserved our attention. The g.2085 T>C

change identified in a subject with high antithrombin levels could potentially modify the

binding of different transcription factors. However, our study revealed that this

polymorphism had no significant effect on antithrombin levels. The low allelic

frequency of this polymorphism (0.027) encourages confirming these results in further

studies including more subjects.

Despite its small size, the dissection of the genetic architecture of the SERPINC1

gene in healthy subjects with extreme antithrombin levels revealed two relevant

findings. A) The genetic variability of this gene is very low. Actually, we only

identified 22 polymorphisms, 1.4 polymorphisms/1,000 bp, a low rate. Particular

interest has the nearly absence of polymorphisms in the coding region of this gene. We

have only identified two linked polymorphisms in exon 5, which in addition did not

modify the aminoacidic sequence (rs5877 and rs5878). A deep review of the literature

and SNPs data bases (Bayston T, Lane D. Imperial College of London. Antithrombin

mutation database.

http://www1.imperial.ac.uk/departmentofmedicine/divisions/experimentalmedicine/hae

matology/coag/antithrombin/; http://www.genecards.org/cgi-

bin/carddisp.pl?gene=SERPINC1&snp=267&search=serpinc1&rf=/home/genecards/cur

rent/website/carddisp.pl#snp) only reveals 19 missense polymorphisms (Table 13),

some of them with low prevalence or associated with antithrombin deficiency.



Discussion

91

Table 13. Missense polymorphisms identified in the SERPINC1 gene.

Chr. position

mRNA pos

dbSNP rs# cluster id MAF dbSNP

allele Protein residue

Amino acid pos*

Clinical interpretation**

173873072 1469 rs5879 T Asn [N] 450

C Asn [N] 450 173873198 1343 rs201411353 0.0005 C Asn [N] 408

T Asn [N] 408 173876599 1326 rs138743710 T Ser [S] 403 Probable-pathogenic

G Ala [A] 403 173876634 1291 rs201541724 A Gln [Q] 391

G Arg [R] 391 173878720 1242 rs149006854 A Met [M] 375 Probable-pathogenic

C Leu [L] 375 173878832 1130 rs5878 0.4675 G Gln [Q] 337

A Gln [Q] 337 173878838 1124 rs192187532 0.0037 A Val [V] 335

G Val [V] 335 173878862 1100 rs5877 0.4528 G Val [V] 327

A Val [V] 327 173878868 1094 rs200857896 G Ala [A] 325

C Ala [A] 325 173878933 1029 rs150507232 C Leu [L] 304

T Leu [L] 304 173878960 1002 rs201381904 0.0005 A Met [M] 295

G Val [V] 295 173878961 1001 rs184508490 0.0005 T Arg [R] 294

C Arg [R] 294 173878985 977 rs139463995 C His [H] 286 Probable-pathogenic

G Gln [Q] 286 173879035 927 rs139275128 T Leu [L] 270

C Leu [L] 270 173879905 868 rs144084678 T Ile [I] 250 Probable-pathogenic

C Thr [T] 250 173879935 838 rs200861147 0.0005 G Ser [S] 240

A Asn [N] 240 173879958 815 rs141292014 C Asp [D] 232

T Asp[D] 232

http://www.ncbi.nlm.nih.gov/nuccore/NT_004487.19?report=graph&m=173873072&v=173873022:173873122&c=3366FF&theme=Details&flip=false&select=null&content=5&color=0&decor=0&layout=0&spacing=0�

http://www.ncbi.nlm.nih.gov/nuccore/NM_000488.3?report=graph&v=1419:1519&content=5&m=1469&dispmax=1&currpage=1�

http://www.ncbi.nlm.nih.gov/projects/SNP/snp_ref.cgi?rs=5879�

http://www.ncbi.nlm.nih.gov/protein/NP_000479.1?report=graph&v=400:500&content=5&m=450�

































































Discussion

92

Chr. position

mRNA pos



Amino acid pos*


173879961 812 rs140452859 0.0014 T Thr [T] 231

C Thr [T] 231 173879976 797 rs2227627 0.0009 T Thr [T] 226

C Thr [T] 226 173879984 789 rs146733468 G Asp [D] 224 Probable-pathogenic

A Asn [N] 224 173880967 713 rs183416252 0.0005 C Tyr [Y] 198

T Tyr [Y] 198 173881032 648 rs143521873 T Cys [C] 177 Probable-pathogenic

C Arg [R] 177 173881122 558 rs2227606 0.0023 G Ala [A] 147 Probable-pathogenic

A Thr [T] 147 173881135 545 rs199525344 0.0005 A Thr [T] 142

C Thr [T] 142 173883748 470 rs200810296 0.0005 A Thr [T] 117

G Thr [T] 117 173883749 469 rs139392083 T Met [M] 117 Probable-pathogenic

C Thr [T] 117 173883788 430 rs200118419 T Val [V] 104 Probable-pathogenic

A Asp [D] 104 173883797 421 rs199895690 0.0005 G Cys [C] 101 Probable-pathogenic

C Ser [S] 101 173883834 384 rs147266200 0.0005 T Cys [C] 89 Pathogenic

C Arg [R] 89 173883987 231 rs145771113 A Asn [N] 38

G Asp [D] 38 173884000 218 rs147676453 A Gln [Q] 33

T His [H] 33

C His [H] 33 173884010 208 rs2227624 0.0018 A Glu [E] 30

T Val [V] 30 173884012 206 rs188274879 0.0005 T Cys [C] 29

C Cys [C] 29 173884014 204 rs147918976 C Arg [R] 29

T Cys [C] 29 173886369 148 rs61736655 0.0005 A Asn [N] 10









































































Discussion

93

Chr. position

mRNA pos



Amino acid pos*


C Thr [T] 10 *: numbering from first Met. **: Clinical interpretation based on identification in patients with antithrombin deficiency (Pathological) or the same residue affected in a patient with antithrombin deficiency, affecting a highly conserved or functional residue (probable pathogenic). Lines in gray: missense polymorphisms Lines in white: silent polymorphisms Data from http://www.ncbi.nlm.nih.gov/sites/varvu?gene=462&rs=267598177 http://www.ncbi.nlm.nih.gov/projects/SNP/snp_ref.cgi?geneId=462

Antithrombin deficiency associated to missense mutations may be caused by

mechanisms common to many other proteins, such as RNA instability. However, the

extraordinary functional and conformational sensitivity of this molecule [79] could also

contribute to explain the low prevalence of missense polymorphisms. Antithrombin, as

a member of the serpin superfamily has to fold into a metastable conformation. Minor

modification affecting the folding of this molecule, particularly those mutations

affecting key structural residues, might impair the folding process and the correct

intracellular pathway. The number of candidate residues is really high due to the

conservation among serpins. Moreover, antithrombin also has functional domains

highly sensitive to even minor mutations. Thus, changes of residues involved in the

interaction with heparin and the latter activation process, residues involved in the

interaction with the target proteases or its translocation process after the proteolytic

cleavage might have deleterious consequences that could impair or abolish the

inhibitory capacity of this molecule. All these data suggest that most changes, even

affecting a single residue, might have significant consequences leading to antithrombin

deficiency. Further genetic studies in patients with thrombosis but not significant

antithrombin deficiency are encouraged to verify if mild mutations with potential

pathological relevance might exist.

B) The genetic variability of SERPINC1 plays a minor role on the interindividual

variability of antithrombin levels in plasma. Only one polymorphism in intron 1

(rs2227589) explains a minor percentage of the interindividual variation (7%).

The low genetic variability of SERPINC1 together with the high heritability of

antithrombin levels [64] support the existence of modulating genes that might be

involved in s variability, and could play a role in the course or severity of thrombotic

disorders [123]. These genes might encode for transcriptional factors, molecules

http://www.ncbi.nlm.nih.gov/sites/varvu?gene=462&rs=267598177�

http://www.ncbi.nlm.nih.gov/projects/SNP/snp_ref.cgi?geneId=462�

Discussion

94

involved in folding or secretion, enzymes implicated in posttranslational modifications,

proteases or miRNAs. An effort to identify these potential modulating genes of

antithrombin has also been done in this study.

2) The first objective has confirmed the low genetic variability of SERPINC1

locus and its minor functional influence on antithrombin levels. Therefore we focused

our search to other genes that could indirectly modulate antithrombin by a multi-stage

approach, the same approach followed by a very recent study that extended the search to

protein C and protein S [124]. The potential identification of modulators of key

anticoagulant genes might help to identify new genetic risk factors for venous

thrombosis. Moreover, this study might also identify new mechanisms or pathways

involved in the folding, secretion, function or clearance of antithrombin, which may

also be extrapolated to other homologous proteins. A similar approach might be

considered for other haemostatic factors, with interindividual variability not fully

explained by the genetic variability of the coding gene, particularly considering the few

studies that have tried to identify modifying genes of haemostatic factors [123].

However, both multi-stage approaches failed to find SNPs associated with

antithrombin levels at genome-wide significance. These negative results, despite the

high heritability of antithrombin levels, strongly suggest that different elements with

potential moderate effect might modulate the levels of this key anticoagulant and

strength the requirement of additional approaches using different experimental studies,

like those used in this objective of our study, to identify antithrombin-modulating genes.

Thus, our replication analysis has evaluated 12 SNPs with milder association with anti-

FXa activity in the GWAS, three rs762057, rs713703 and rs240082 affecting LARGE.

One SNP, rs762057, maintained the association with anti-FXa activity in the validation

cohort, and interestingly, in both cohorts the polymorphic allele associated with a

modest increase of anti-FXa activity in plasma (2-4%). This result also explains the

absence of thrombotic relevance for this polymorphism in the case-control study.

Moreover, the LARGE H2 haplotype defined by rs762057, rs713703 and rs240082, also

associated with a modest increase of anti-FXa activity. The relevance of these three

LARGE SNPs on the heritability of anti-FXa-levels was minor: 4.3%, 4.9% and 4.0%

for rs762057, rs713703 and rs240082, respectively, but rose to 7.8% when considering

the three SNPs together. It is possible that other polymorphisms not included in the

chip, haplotypes or rare mutations of LARGE might have stronger functional

Discussion

95

consequences. Unfortunately, the name of this gene reflects its length and genetic

variability. It expands more than 756,000 bp and contains more than 7,790

polymorphisms, a size and genetic variability that may complicate the studies required

to dissect the genetic architecture of this gene and to evaluate its potential functional

and pathological relevance.

As the GWAS approach and the validation study identified LARGE as a candidate

antithrombin-modulating gene based only on association data, additional experimental

evidences were required to sustain a potential role of LARGE on the indirect regulation

of the levels of this anticoagulant. Thus silencing experiments confirmed a role for

LARGE modulating antithrombin levels. Moreover, these results may also open new

mechanisms or pathways involved in the folding, secretion, function or clearance of this

important anticoagulant, which may also be extrapolated to other homologous proteins.

Additionally, our study also opens new attractive roles for LARGE, a protein

largely unknown. The activity of LARGE is required for the right function alpha-

dystroglycan (α-DG) as mutations in the LARGE gene have been identified in

congenital muscular dystrophy patients with brain abnormalities [125]. Early evidences

suggested that LARGE participates in the formation of a phosphoryl glycan branch on

O-linked mannose or it modifies complex N- and mucin O-glycans [126,127]. Indeed,

site-directed mutagenesis of the transferase-like domains abolishes its glycosylation

capability, suggesting that LARGE may function as a glycosyltransferase [126].

Despite LARGE plays a critical role in biosynthesis of the functional O-glycans of α-

dystroglycan (α-DG) [128], it is also described that its over expression competes to

modify GlcNAc terminals with Gal to generate the functional glycans not only in O-

linked but also in N-glycans in α-DG [129]. Moreover, a recent study over expressing

LARGE suggested that this protein can also mediate phosphoryl glycosylation on N-

linked glycans of non-α-DG proteins [130]. Finally, an excellent and recent study

demonstrated that LARGE could act as a bifunctional glycosyltransferase, with both

xylosyltransferase and glucuronyltransferase activities, which produced repeating units

of [–3-xylose–α1,3–glucuronic acid-β1–] [131]. How could LARGE modulate

antithrombin levels? Since reduced expression of LARGE did not affect the expression

of SERPINC1, we can rule out an indirect role of LARGE on the transcriptional

regulation of antithrombin. According to our results, a direct effect on the glycomic

features of antithrombin might also be discarded. The reduced expression of LARGE

seems to down-regulate the secretion of antithrombin, without significant intracellular

Discussion

96

accumulation, probably reflecting a degradation of abnormal folding proteins [132].

These data together with the impaired secretion of other proteins (α1-antitrypsin,

prothrombin or transferrin) observed under silencing of LARGE encouraged us to

suggest a new function for LARGE in intracellular folding and/or secretion. The fact

that the main affected protein among all tested is antithrombin, a protein with an heparin

binding domain [133], together with the fact that the glycan produced by LARGE

resembles heparin-heparan sulfate (HS) and chondroitin-dermatan sulfate (CS-DS)

glycosaminoglycans (GAGs) [131], make attractive this hypothesis. Further studies are

required to verify this hypothesis and to define the exact mechanism involving LARGE

on the folding, secretion and degradation pathways of glycoproteins, particularly

antithrombin, and to determine the final effect on the haemostatic equilibrium, as

LARGE might also reduce the secretion of prothrombotic proteins such as prothrombin.

In conclusion, through a hypothesis free GWAS approach followed by functional

experiments, we have identified the first modulating gene of antithrombin: LARGE.

This is a bifunctional glycosyltransferase that produces a glycan that resembles heparin-

HS– and CS-DS–GAGs linear polysaccharides. Our study also found that under

expression of LARGE also impairs the secretion of many other proteins, suggesting that

LARGE could also have additional roles, including a potential role in the appropriate

folding and/or secretion pathways of multiple proteins.

3) In the deep seeking to find new genetic elements involved in the risk of

thrombosis, rare mutations affecting different elements of the haemostatic system are

emerging candidates [8,9]. The identification of LARGE as a potential antithrombin

modulating gene opens new perspectives in thrombophilia. Accordingly, mutations

affecting other genes able to influence antithrombin levels or the whole haemostatic

system may be alternative thrombophilic candidates. In the third objective of the present

study, we identified a mechanism leading to severe antithrombin deficiency with a

significant risk of thrombosis. The analysis of patients with thrombosis and

antithrombin deficiency but without mutations or deletions in the SERPINC1 gene was

crucial to identify these new mechanisms.

We firstly focused on the analysis of antithrombin in a thrombophilic patient with

severe deficiency and early and recurrent venous thrombosis. This study allowed us to

identify a particular thrombophilic defect that, affecting the N-glycosylation pathway,

disturbs this key anticoagulant serpin. This patient (P27) developed four severe

Discussion

97

thrombotic episodes and antithrombin deficiency was the single thrombophilic factor

identified. The initial identification of abnormal antithrombin glycoforms by

electrophoretic analysis followed by different biochemical studies, a deeply

glycoprotein study and a PMM2 molecular characterization has provided the final

accurate diagnosis of Congenital Disorder of Glycosylation type Ia (CDG-Ia, also called

PMM2-CDG).

Congenital Disorders of Glycosylation (CDG) result from defects in the assembly,

transfer, and processing of N-linked oligosaccharides. Since 1980, when the first N-

glycosylation defect was described and characterized [134] over 45 types of CDGs have

been recognized. Although very rare, CDGs are the most rapidly expanding group

among the currently known 3,000 monogenetic inherited diseases. All known CDGs

have a recessive inheritance except EXT1/EXT2-CDG which is dominantly inherited

and MGAT1-CDG which is X linked [110].

CDG-Ia (PMM2-CDG, OMIM#212065) is by far the most prevalent protein N-

glycosylation defect with more than 700 affected individuals worldwide. The

prevalence could be as high as 1:20,000 [135]. Phosphomannomutase (PMM) 2

deficiency is a (cytosolic) defect in the second step of the mannose pathway

(transforming mannose 6-phosphate into mannose 1-phosphate), which normally leads

to the synthesis of GDP-mannose. This nucleotide sugar is the donor of the mannoses

used in the endoplasmic reticulum to assemble the dolichol-pyrophosphate

oligosaccharide precursor (Figure 41). Deficiency of GDP-mannose causes

hypoglycosylation of numerous glycoproteins, including serum proteins, lysosomal

enzymes, and membranous glycoproteins [136].

Discussion

98

Figure 41: N-glycosylation of proteins and its defects. The figure shows the enzymatic steps taking

place in the cytosol and in the endoplasmic reticulum. Enzymatic defects in any of these steps leads to

CDG (congenital disorder of glycosylation). N-glycosylation also involves steps taking place in the Golgi

apparatus (not shown), which may also be defective.

The clinical spectrum associated to CDGs in general and PMM2-CDG in

particular, is very large. The nervous system is affected in all patients, and most other

organs are involved in a variable way. The neurological picture comprises alternating

internal strabism and other abnormal eye movements, axial hypotonia, psychomotor

retardation, ataxia, and hyporeflexia. After infancy, symptoms include retinitis

pigmentosa, often stroke-like episodes, and sometimes epilepsy. During the first year(s)

of life, there are variable feeding problems such as anorexia, vomiting, and diarrhea.

These can lead to severe failure to thrive. Other features include variable dysmorphy

(large hypoplastic/dysplastic ears, abnormal subcutaneous adipose tissue distribution),

hepatomegaly, skeletal abnormalities, and hypogonadism. Some infants develop

pericardial effusion and/or cardiomyopathy. At the other end of the spectrum are

patients with a very mild phenotype (no dysmorphy, mild psychomotor retardation).

Patients often have an extroverted and happy appearance. There is an increased

mortality in the first years of life due to vital organ involvement or severe infection.

Approximately 90 mutations have been identified in the PMM2 gene. The mutation

Discussion

99

leading to the p.Arg141His substitution is present in about 75% of the alleles of Central

European patients [112].

The haemostatic system in these patients is also relatively impaired. Up to 55% of

PMM2-CDG patients show acute events such as ‘‘stroke-like’’ episodes, ischemic

cerebral vascular accidents, and at least 10 cases with either venous or arterial

thrombosis have been described [137]. Acute vascular events occurred in patients

younger than 15 years, especially during fever and prolonged immobilization [137].

These accidents can be recurrent in the same patient [137]. Only few studies have

evaluated haemostatic factors [137,138] and platelets [139] in PMM2-CDG with the

aim to identify the mechanism(s) involved in the development of these vascular events.

PMM2-CDG patients consistently have antithrombin deficiency [137,138] with

abnormal hypoglycosylated antithrombin molecules [137]. This might contribute to

increase the risk of thrombotic events described in these patients. However, there are no

differences in antithrombin levels between patients with and without vascular events

[137].

P27 is the first report of a PMM2-CDG patient with documented and recurrent

venous thrombosis. Moreover, the patient suffered from thrombotic events without a

clear triggering factor such as fever, catheterization or immobilization. Finally, two

recurrent thrombotic events appeared under oral anticoagulant therapy. This unusual

clinical severity suggests that additional haemostatic or vascular defects than those

described in PMM2-CDG patients (or additional unknown thrombophilic anomalies)

might be present in this patient.

This case also highlights another important conclusion concerning the diagnostic

procedure in patients with antithrombin deficiency. Direct sequencing and MLPA

analysis of the SERPINC1 gene in patients with thrombosis and antithrombin deficiency

is highly effective, but these studies failed to find any genetic modification in up to 10%

of these patients [73,74]. Our study suggests that a diagnosis of PMM2-CDG might be

suspected among the cases, particularly in patients with antithrombin deficiency without

a family history of thrombosis, as well as in cases with a lower molecular weight

plasma antithrombin.

Finally, this new thrombophilic factor could be more frequent than expected. The

electrophoretic analysis of antithrombin in 25 adult patients with thrombosis and

moderate antithrombin deficiency but no mutation in the SERPINC1 gene has identified

four patients with smaller forms of antithrombin also caused by defects in N-

Discussion

100

glycosylation. Further studies in these patients confirmed that this is caused by a non-

specific defect, as other hepatic glycoproteins (α1-antitrypsin and transferrin) were also

affected according to electrophoresis, HPLC and Q-TOF methods. Moreover, molecular

analysis of PMM2 gene identified heterozygous mutations in three of them. The

importance of these findings is not only restricted to the high frequency of this new

disorder, but also in the clinical features of these patients and the pathological

mechanism.

The patient with a biochemical phenotype closer to the PMM2-CDG disorder,

P15, carried the PMM2 c.422G>A missense mutation responsible for the change

p.Arg141His, the most frequent alteration in PMM2-CDG patients. Around 37% of

PMM2-CDG patients are heterozygous for this mutation, which has never been

observed in the homozygous state, not even in North Europe where this alteration

reaches the frequency of a polymorphism (1/72 in Dutch and Danish populations)

[140,141]. This mutation does not abolish but significantly reduces the

phosphomannomutase activity (to a very low residual activity, <1%) [142]. In P15, the

p.Arg141His mutation was of paternal origin, but as described for parents of PMM2-

CDG patients, heterozygous carriers have neither biochemical phenotype nor clinical

symptoms. P15 also had a Darier syndrome which was inherited by his mother. Our first

hypothesis considered that the combination of a mutation in the ATP2A2 gene (the gene

affected in patients with Darier disease) and in the PMM2 gene might result in a CDG-

like disorder. However, it was discarded by several evidences: 1) no reduced

glycosylation of proteins, reduced antithrombin levels or thrombosis has been identified

in patients with Darier syndrome, 2) the use of Ca2+-pump inhibitors did not modify

glycosylation of proteins in HepG2 cells, and more importantly, 3) the variable

laboratory data of P15 during 5 years of follow up, 4) the absence of neurological

symptoms in P15, and 5) any other CDG-like patient identified in our study had a

Darier disease. Moreover, all these data argued against the combination of two

congenital defects to explain the CDG-like disorder and strongly suggest a role for an

acquired factor in combination with a PMM2 defect.

Two additional CDG-like patients P16 and P2 carried PMM2 intronic mutations

(IVS4 +21 G>C and IVS2 -13 G>A, respectively). In silico studies revealed these

mutations could generate new branch-sites. Branch-site sequence is generally located

between 10 and 50 nucleotides upstream of the 3’consensus splice site (AG).

Interestingly, both mutations are located within these distances. Although, mutations in

Discussion

101

branch sites seems to be a relatively rare class of splice mutations several instances of

point mutations within the branch point sequence have been published, for example one

in the COL5A1 gene associated with type II Ehlers–Danlos syndrome [143] and two

others in a DNA repair gene in xeroderma pigmentosum [144]. Therefore branch-site

mutations are possibly often undiagnosed [145].

These PMM2 intronic mutations were not identified in healthy people, but their

functional effect has not been elucidated yet. Further studies of the PMM2 activity in

fibroblast or PMM2 transcript analysis, and/or exon trap studies should be performed to

better characterize the functional consequences of these intronic changes. However, we

can expect a lower functional effect than that of p.Arg141His according to the reduced

levels of underglycosylated proteins detected in carriers of these intronic mutations.

Finally, the last CDG-like patient, P22, has no PMM2 mutations. It is possible that

mutations affecting other regulatory regions of PMM2 gene might be present in this

patient, but we can’t discard a mutation in other gene involved in CDG disorders.

Indeed, there are 45 genes identified so far associated with CDGs [136], a number that

will grow since there are patients with clinical and biochemical features of CDG

without mutation in any of these genes [146].

Interestingly, all four CDG-like patients described high alcohol intake, and

hepatic transaminases moderately elevated. Severe alcoholism has been previously

associated with glycosylation defects [147]. Briefly, prolonged exposure to ethanol

generates reactive oxygen species (ROS) [148]. Dolichol (an isoprene derivate precursor

of the early steps of N-glycosylation pathway, Figure 41) is particularly vulnerable to

the lipoperoxidative effects of ethanol-induced ROS. Indeed, Cottalasso et al.

demonstrated that ethanol treatment causes a profound reduction in the dolichol

(phosphate) content of rat liver microsomes and Golgi apparatus [148]. A sufficient

level of microsomal dolichol-P is a prerequisite for the initiation of N-linked

glycosylation (Figure 41). Thus, ethanol exposure may seriously impede glycoprotein

synthesis and secretion. Ethanol also affects mannosyltransferase activity, an enzyme

mediating mannosylation of dolichol-P during the build-up of the lipid-linked

oligosaccharide (LLO) [148] (Figure 41). Furthermore, sialyltransferase and

galactosyltransferase are inhibited, while sialidase activity is stimulated in response to

ethanol administration [148,149]. These latter effects of ethanol, causing disruption of

terminal sialylation during the second phase of N-linked glycosylation in the Golgi

apparatus, are probably of minor importance, considering that carbohydrate-deficient

Discussion

102

transferrin (CDT) –the best indicator of chronic alcohol abuse currently available-

mainly consists of di- and asialotransferrin (type I TIEF pattern), indicating defects in

LLO assembly during the first phase of N-linked glycosylation in the Endoplasmic

Reticulum [149,150]. Moreover, in addition to the consequences on dolichol depletion,

degradation of vital membrane constituents by ethanol exposure affects plasma

membrane integrity and vesicular trafficking [148,151]. Finally, alcohol inhibits

gluconeogenesis pathway [152] and in consequence could interfere the early steps of the

N-glycosylation pathway because of the glucose content reduction. Thus, reduced levels

of fructose-6-phosphate in alcoholic patients might be compensated by increased

synthesis through the mannose-6-phosphate isomerase (MPI) activity, reducing the

synthesis of mannose-1 phosphate, thereby aggravating a single PMM2 defect as shown

in Figure 42. Actually, the most recent progress on PMM2-CDG treatment consists on

mannose supplementation and MPI inhibition [153].

Figure 42. Potential interference of alcohol on the N-glycosylation metabolic pathway. HK:

Hexokinase; MPI: Mannose-6 phosphate isomerase; PMM2: phosphomannomutase; Man-6-P: mannose-

6-phosphate; Fru-6-P: fructose-6-phosphate; Man-1-P: mannose-1-phosphate.

However, it is important to point out that isolated alcohol abuse, even at higher

levels than that found in CDG-like patients does not cause the CDG electrophoretic or

HLPC pattern. We have studied plasma samples from one patient with severe

alcoholism (218 g/day). The electrophoretic pattern of α1-antitrypsin was normal

Discussion

103

(Figure 43) and HPLC analysis of transferrin glycoforms only showed a moderate

increase of the asialotransferrin content, while di- or trisilaotransferrin were normal

(Figure 44).

Figure 43. Electrophoretic pattern of α1-Antitrypsin of an alcoholic patient comparing with a

control, PMM2-CDG and CDG-like patients. Alcohol intake is indicated below.

Figure 44. HPLC profile of plasma transferrin glycoforms from an alcoholic patient.

All these results strongly suggest that CDG-like disorders might be the result of

the combination of a moderate congenital defect affecting the N-glycosylation pathway

(mainly PMM2) and an acquired factor also affecting this pathway, being alcohol an

excellent candidate. Further studies using animal models will be required to confirm

that this combination produces CDG-like disorders. It is also possible that other

Discussion

104

combinations of different genes and acquired factors might also render CDG or CDG-

like disorders.

This study also aimed to evaluate the incidence of this new thrombophilic disorder

in different thromboembolic diseases. Our preliminary epidemiological studies aiming

to determine the incidence in thrombotic diseases of CDG-like disorders suggest a

relatively high incidence in both venous and arterial thrombosis. It is true that the high

incidence of CDG-like disorders among patients with moderate antithrombin deficiency

(4/25: 16%) make us to expect higher prevalence of CDG-like in thrombotic disorders

than that found in our study, particularly in DVT and specially in the north of Europe,

where the p.Arg141His is a relatively common polymorphism. However, to fully

understand the thrombotic relevance of these new disorders, it would be necessary to

perform a multivariate analysis including these acquired factors that exacerbate the

PMM2 defect, at least high alcohol intake. Moreover, we can’t forget that the abnormal

glycosylation pattern used to identify the CDG-like disorder is highly variable (the best

example is P15 during the 5 years of study) and depends on the combination with the

acquired factor. It is therefore possible that our epidemiological study, performed with

plasma samples of retrospectively recruited patients had underestimated the incidence

of these disorders. Further studies using samples obtained during the acute event might

answer this question.

Finally, it is necessary to investigate the mechanism responsible for the high risk

of thrombosis described in CDG and CDG-like patients. The deficiency of antithrombin

is an excellent candidate, but we do not know if this is cause or consequence of the

disorder. Both anticoagulant and procoagulant hepatic proteins are affected in these

patients (antithrombin and prothrombin are the best examples). Trying to identify

further defects on haemostatic elements we studied platelet glycoproteins. Surprisingly,

we observed no major modifications on all proteins tested, even using high sensitive

methods such as proteomic studies. These results sustain that the impaired glycosylation

pathway of PMM2-CDG and CDG-like patients has different consequences depending

on the protein and particularly the cell type. Our results suggest that cells with a high

synthesis and secretory rate like liver are potentially more affected by these disorders.

Further studies are therefore required to answer how disorders of glycosylation affect

the haemostatic system and to identify the mechanism underlying the increased risk of

thrombosis of these patients.

Conclusions

105

CONCLUSIONS

Conclusions

106

Conclusions

107

The result of this study reveals a fascinating complex heterogeneity of new

mechanisms affecting the levels of antithrombin a key anticoagulant serpin, with

potential pathological consequences.

1) We identified the first mutations in the potential promoter of SERPINC1 gene that

reduced the levels of plasma antithrombin. This region must be included in

molecular analysis aiming to identify the genetic defect involved in congenital


2) SERPINC1 locus has low genetic variability with minor influence in the

heterogeneity of the plasma antithrombin in the general population.

3) We have identified the first antithrombin modulating gene. The reduced levels or

activity of LARGE, potentially through its xylosyltransferase and

glucuronyltransferase activities that produce a heparin-like glycosaminoglycan,

impair the secretion of antithrombin, a heparin-binding protein.

4) We diagnose a PMM2-CDG disorder in a young patient with recurrent venous

thrombosis and antithrombin deficiency.

5) We identified a new CDG-like disorder based on the similar electrophoretic,

HPLC or Q-TOF pattern of glycoforms of hepatic proteins than of PMM2-CDG

patients, probably as the result of combination of a PMM2 mutation and

alcoholism.

6) CDG-like, as PMM2-CDG, may be considered as new thrombophilic disorder and

must be studied in patients with venous thrombosis, particularly in those with

moderate or severe antithrombin deficiency and no mutation or deletion in the

SERPINC1 gene.

Conclusions

108

Bibliography

109

BIBLIOGRAPHY

Bibliography

110

Bibliography

111

Reference List

1. Lippi G, Favaloro EJ, Franchini M, Guidi GC. Milestones and perspectives in coagulation and hemostasis. Semin Thromb Hemost 2009; 35; 9-22.

2. Egeberg O. Inherited antithrombin deficiency causing thrombophilia. Thromb Diath Haemorrh 1965; 13; 516-30.

3. Souto JC, Almasy L, Borrell M, Blanco-Vaca F, Mateo J, Soria JM, Coll I, Felices R, Stone W, Fontcuberta J, Blangero J. Genetic susceptibility to thrombosis and its relationship to physiological risk factors: the GAIT study. Genetic Analysis of Idiopathic Thrombophilia. Am J Hum Genet 2000; 67; 1452-9.

4. Bertina RM. Genetic approach to thrombophilia. Thromb Haemost 2001; 86; 92-103.

5. Gohil R, Peck G, Sharma P. The genetics of venous thromboembolism. A meta-analysis involving approximately 120,000 cases and 180,000 controls. Thromb Haemost 2009; 102; 360-70.

6. Germain M, Saut N, Greliche N, Dina C, Lambert JC, Perret C, Cohen W, Oudot-Mellakh T, Antoni G, Alessi MC, Zelenika D, Cambien F, Tiret L, Bertrand M, Dupuy AM, Letenneur L, Lathrop M, Emmerich J, Amouyel P, Tregouet DA, et al. Genetics of venous thrombosis: insights from a new genome wide association study. PLoS One 2011; 6; e25581.

7. Heit JA, Armasu SM, Asmann YW, Cunningham JM, Matsumoto ME, Petterson TM, de Andrade M. A genome-wide association study of venous thromboembolism identifies risk variants in chromosomes 1q24.2 and 9q. J Thromb Haemost 2012; 10; 1521-31.

8. Miyawaki Y, Suzuki A, Fujita J, Maki A, Okuyama E, Murata M, Takagi A, Murate T, Kunishima S, Sakai M, Okamoto K, Matsushita T, Naoe T, Saito H, Kojima T. Thrombosis from a prothrombin mutation conveying antithrombin resistance. N Engl J Med 2012; 366; 2390-6.

9. Simioni P, Tormene D, Tognin G, Gavasso S, Bulato C, Iacobelli NP, Finn JD, Spiezia L, Radu C, Arruda VR. X-linked thrombophilia with a mutant factor IX (factor IX Padua). N Engl J Med 2009; 361; 1671-5.

10. Law RH, Zhang Q, McGowan S, Buckle AM, Silverman GA, Wong W, Rosado CJ, Langendorf CG, Pike RN, Bird PI, Whisstock JC. An overview of the serpin superfamily. Genome Biol 2006; 7; 216.

11. Hunt LT, Dayhoff MO. A surprising new protein superfamily containing ovalbumin, antithrombin-III, and alpha 1-proteinase inhibitor. Biochem Biophys Res Commun 1980; 95; 864-71.

12. Bird PI. Serpins and regulation of cell death. Results Probl Cell Differ 1998; 24; 63-89.

Bibliography

112

13. Irving JA, Pike RN, Lesk AM, Whisstock JC. Phylogeny of the serpin superfamily: implications of patterns of amino acid conservation for structure and function. Genome Res 2000; 10; 1845-64.

14. Whisstock JC, Bottomley SP. Molecular gymnastics: serpin structure, folding and misfolding. Curr Opin Struct Biol 2006; 16; 761-8.

15. Rau JC, Beaulieu LM, Huntington JA, Church FC. Serpins in thrombosis, hemostasis and fibrinolysis. J Thromb Haemost 2007; 5 Suppl 1; 102-15.

16. Silverman GA, Bird PI, Carrell RW, Church FC, Coughlin PB, Gettins PG, Irving JA, Lomas DA, Luke CJ, Moyer RW, Pemberton PA, Remold-O'Donnell E, Salvesen GS, Travis J, Whisstock JC. The serpins are an expanding superfamily of structurally similar but functionally diverse proteins. Evolution, mechanism of inhibition, novel functions, and a revised nomenclature. J Biol Chem 2001; 276; 33293-6.

17. Bode W, Huber R. Structural basis of the endoproteinase-protein inhibitor interaction. Biochim Biophys Acta 2000; 1477; 241-52.

18. Cho YL, Chae YK, Jung CH, Kim MJ, Na YR, Kim YH, Kang SJ, Im H. The native metastability and misfolding of serine protease inhibitors. Protein Pept Lett 2005; 12; 477-81.

19. Huntington JA, Read RJ, Carrell RW. Structure of a serpin-protease complex shows inhibition by deformation. Nature 2000; 407; 923-6.

20. Belorgey D, Hagglof P, Karlsson-Li S, Lomas DA. Protein misfolding and the serpinopathies. Prion 2007; 1; 15-20.

21. Davies MJ, Lomas DA. The molecular aetiology of the serpinopathies. Int J Biochem Cell Biol 2008; 40; 1273-86.

22. Cugno M, Zanichelli A, Foieni F, Caccia S, Cicardi M. C1-inhibitor deficiency and angioedema: molecular mechanisms and clinical progress. Trends Mol Med 2009; 15; 69-78.

23. Quinsey NS, Greedy AL, Bottomley SP, Whisstock JC, Pike RN. Antithrombin: in control of coagulation. Int J Biochem Cell Biol 2004; 36; 386-9.

24. Martinez-Martinez I, Ordonez A, Navarro-Fernandez J, Perez-Lara A, Gutierrez-Gallego R, Giraldo R, Martinez C, Llop E, Vicente V, Corral J. Antithrombin Murcia (K241E) causing antithrombin deficiency: a possible role for altered glycosylation. Haematologica 2010; 95; 1358-65.

25. Picard V, Ersdal-Badju E, Bock SC. Partial glycosylation of antithrombin III asparagine-135 is caused by the serine in the third position of its N-glycosylation consensus sequence and is responsible for production of the beta-antithrombin III isoform with enhanced heparin affinity. Biochemistry 1995; 34; 8433-40.

Bibliography

113

26. Olson ST, Chuang YJ. Heparin activates antithrombin anticoagulant function by generating new interaction sites (exosites) for blood clotting proteinases. Trends Cardiovasc Med 2002; 12; 331-8.

27. McCoy AJ, Pei XY, Skinner R, Abrahams JP, Carrell RW. Structure of beta-antithrombin and the effect of glycosylation on antithrombin's heparin affinity and activity. J Mol Biol 2003; 326; 823-33.

28. Swedenborg J. The mechanisms of action of alpha- and beta-isoforms of antithrombin. Blood Coagul Fibrinolysis 1998; 9 Suppl 3; S7-10.

29. Luxembourg B, Delev D, Geisen C, Spannagl M, Krause M, Miesbach W, Heller C, Bergmann F, Schmeink U, Grossmann R, Lindhoff-Last E, Seifried E, Oldenburg J, Pavlova A. Molecular basis of antithrombin deficiency. Thromb Haemost 2011; 105; 635-46.

30. Berry LR, Thong B, Chan AK. Comparison of recombinant and plasma-derived antithrombin biodistribution in a rabbit model. Thromb Haemost 2009; 102; 302-8.

31. Carlson TH, Atencio AC, Simon TL. Comparison of the behaviour in vivo of two molecular forms of antithrombin III. Biochem J 1985; 225; 557-64.

32. Frebelius S, Isaksson S, Swedenborg J. Thrombin inhibition by antithrombin III on the subendothelium is explained by the isoform AT beta. Arterioscler Thromb Vasc Biol 1996; 16; 1292-7.

33. Kamp P, Strathmann A, Ragg H. Heparin cofactor II, antithrombin-beta and their complexes with thrombin in human tissues. Thromb Res 2001; 101; 483-91.

34. Swedenborg J. The mechanisms of action of alpha- and beta-isoforms of antithrombin. Blood Coagul Fibrinolysis 1998; 9 Suppl 3; S7-10.

35. Witmer MR, Hatton MW. Antithrombin III-beta associates more readily than antithrombin III-alpha with uninjured and de-endothelialized aortic wall in vitro and in vivo. Arterioscler Thromb 1991; 11; 530-9.

36. Stein PE, Carrell RW. What do dysfunctional serpins tell us about molecular mobility and disease? Nat Struct Biol 1995; 2; 96-113.

37. Whisstock J, Skinner R, Lesk AM. An atlas of serpin conformations. Trends Biochem Sci 1998; 23; 63-7.

38. Gils A, Declerck PJ. Structure-function relationships in serpins: current concepts and controversies. Thromb Haemost 1998; 80; 531-41.

39. Wardell MR, Abrahams JP, Bruce D, Skinner R, Leslie AG. Crystallization and preliminary X-ray diffraction analysis of two conformations of intact human antithrombin. J Mol Biol 1993; 234; 1253-8.

Bibliography

114

40. Johnson DJ, Langdown J, Li W, Luis SA, Baglin TP, Huntington JA. Crystal structure of monomeric native antithrombin reveals a novel reactive center loop conformation. J Biol Chem 2006; 281; 35478-86.

41. Carrell RW, Stein PE, Fermi G, Wardell MR. Biological implications of a 3 A structure of dimeric antithrombin. Structure 1994; 2; 257-70.

42. Jin L, Abrahams JP, Skinner R, Petitou M, Pike RN, Carrell RW. The anticoagulant activation of antithrombin by heparin. Proc Natl Acad Sci U S A 1997; 94; 14683-8.

43. Olson ST, Bjork I, Sheffer R, Craig PA, Shore JD, Choay J. Role of the antithrombin-binding pentasaccharide in heparin acceleration of antithrombin-proteinase reactions. Resolution of the antithrombin conformational change contribution to heparin rate enhancement. J Biol Chem 1992; 267; 12528-38.

44. Huntington JA. Mechanisms of glycosaminoglycan activation of the serpins in hemostasis. J Thromb Haemost 2003; 1; 1535-49.

45. Langdown J, Belzar KJ, Savory WJ, Baglin TP, Huntington JA. The critical role of hinge-region expulsion in the induced-fit heparin binding mechanism of antithrombin. J Mol Biol 2009; 386; 1278-89.

46. Petitou M, Duchaussoy P, Herbert JM, Duc G, El HM, Branellec JF, Donat F, Necciari J, Cariou R, Bouthier J, Garrigou E. The synthetic pentasaccharide fondaparinux: first in the class of antithrombotic agents that selectively inhibit coagulation factor Xa. Semin Thromb Hemost 2002; 28; 393-402.

47. Izaguirre G, Rezaie AR, Olson ST. Engineering functional antithrombin exosites in alpha1-proteinase inhibitor that specifically promote the inhibition of factor Xa and factor IXa. J Biol Chem 2009; 284; 1550-8.

48. Martinez-Martinez I, Ordonez A, Pedersen S, de la Morena-Barrio ME, Navarro-Fernandez J, Kristensen SR, Minano A, Padilla J, Vicente V, Corral J. Heparin affinity of factor VIIa: implications on the physiological inhibition by antithrombin and clearance of recombinant factor VIIa. Thromb Res 2011; 127; 154-60.

49. Li W, Johnson DJ, Esmon CT, Huntington JA. Structure of the antithrombin-thrombin-heparin ternary complex reveals the antithrombotic mechanism of heparin. Nat Struct Mol Biol 2004; 11; 857-62.

50. Pike RN, Buckle AM, le Bonniec BF, Church FC. Control of the coagulation system by serpins. Getting by with a little help from glycosaminoglycans. FEBS J 2005; 272; 4842-51.

51. Im H, Ahn HY, Yu MH. Bypassing the kinetic trap of serpin protein folding by loop extension. Protein Sci 2000; 9; 1497-502.

52. Corral J, Rivera J, Martinez C, Gonzalez-Conejero R, Minano A, Vicente V. Detection of conformational transformation of antithrombin in blood with

Bibliography

115

crossed immunoelectrophoresis: new application for a classical method. J Lab Clin Med 2003; 142; 298-305.

53. Mushunje A, Evans G, Brennan SO, Carrell RW, Zhou A. Latent antithrombin and its detection, formation and turnover in the circulation. J Thromb Haemost 2004; 2; 2170-7.

54. O'Reilly MS, Pirie-Shepherd S, Lane WS, Folkman J. Antiangiogenic activity of the cleaved conformation of the serpin antithrombin. Science 1999; 285; 1926-8.

55. Bae JS, Rezaie AR. Mutagenesis studies toward understanding the intracellular signaling mechanism of antithrombin. J Thromb Haemost 2009; 7; 803-10.

56. Guerrero JA, Teruel R, Martinez C, Arcas I, Martinez-Martinez I, de la Morena-Barrio ME, Vicente V, Corral J. Protective role of antithrombin in mouse models of liver injury. J Hepatol 2012;

57. Lane DA, Bayston T, Olds RJ, Fitches AC, Cooper DN, Millar DS, Jochmans K, Perry DJ, Okajima K, Thein SL, Emmerich J. Antithrombin mutation database: 2nd (1997) update. For the Plasma Coagulation Inhibitors Subcommittee of the Scientific and Standardization Committee of the International Society on Thrombosis and Haemostasis. Thromb Haemost 1997; 77; 197-211.

58. Olds RJ, Lane DA, Chowdhury V, De Stefano V, Leone G, Thein SL. Complete nucleotide sequence of the antithrombin gene: evidence for homologous recombination causing thrombophilia. Biochemistry 1993; 32; 4216-24.

59. Bock SC, Levitan DJ. Characterization of an unusual DNA length polymorphism 5' to the human antithrombin III gene. Nucleic Acids Res 1983; 11; 8569-82.

60. Prochownik EV. Relationship between an enhancer element in the human antithrombin III gene and an immunoglobulin light-chain gene enhancer. Nature 1985; 316; 845-8.

61. Fernandez-Rachubinski FA, Weiner JH, Blajchman MA. Regions flanking exon 1 regulate constitutive expression of the human antithrombin gene. J Biol Chem 1996; 271; 29502-12.

62. Ochoa A, Brunel F, Mendelzon D, Cohen GN, Zakin MM. Different liver nuclear proteins binds to similar DNA sequences in the 5' flanking regions of three hepatic genes. Nucleic Acids Res 1989; 17; 119-33.

63. Sheffield WP, Wu YI, Blajchman MA. Antithrombin: Structure and Function. In: High KA, Roberts HR, editors. Molecular Basis of Thrombosis and Hemostasis.New York: Marcel Dekker, Inc.; 1995. p. 355-77.

64. Souto JC, Almasy L, Borrell M, Gari M, Martinez E, Mateo J, Stone WH, Blangero J, Fontcuberta J. Genetic determinants of hemostasis phenotypes in Spanish families. Circulation 2000; 101; 1546-51.

Bibliography

116

65. Anton AI, Teruel R, Corral J, Minano A, Martinez-Martinez I, Ordonez A, Vicente V, Sanchez-Vega B. Functional consequences of the prothrombotic SERPINC1 rs2227589 polymorphism on antithrombin levels. Haematologica 2009; 94; 589-92.

66. Hernandez-Espinosa D, Ordonez A, Minano A, Martinez-Martinez I, Vicente V, Corral J. Hyperglycaemia impairs antithrombin secretion: Possible contribution to the thrombotic risk of diabetes. Thromb Res 2009;

67. Hernandez-Espinosa D, Mota R, Minano A, Ordonez A, Yelamos J, Vicente V, Corral J. In vivo effects of hyperthermia on the functional and conformational characteristics of antithrombin. J Thromb Haemost 2007; 5; 963-70.

68. Hernandez-Espinosa D, Minano A, Martinez C, Perez-Ceballos E, Heras I, Fuster JL, Vicente V, Corral J. L-asparaginase-induced antithrombin type I deficiency: implications for conformational diseases. Am J Pathol 2006; 169; 142-53.

69. Hernandez-Espinosa D, Ayala I, Castells MT, Garcia-Perez B, Martin-Castillo A, Minano A, Arcas I, Vicente V, Corral J. Intracellular retention of hepatic serpins caused by severe hyperlipidemia. Liver Int 2006; 26; 708-15.

70. Bucciarelli P, Passamonti SM, Biguzzi E, Gianniello F, Franchi F, Mannucci PM, Martinelli I. Low borderline plasma levels of antithrombin, protein C and protein S are risk factors for venous thromboembolism. J Thromb Haemost 2012;

71. Ishiguro K, Kojima T, Kadomatsu K, Nakayama Y, Takagi A, Suzuki M, Takeda N, Ito M, Yamamoto K, Matsushita T, Kusugami K, Muramatsu T, Saito H. Complete antithrombin deficiency in mice results in embryonic lethality. J Clin Invest 2000; 106; 873-8.

72. Lijfering WM, Brouwer JL, Veeger NJ, Bank I, Coppens M, Middeldorp S, Hamulyak K, Prins MH, Buller HR, Van Der Meer J. Selective testing for thrombophilia in patients with first venous thrombosis: results from a retrospective family cohort study on absolute thrombotic risk for currently known thrombophilic defects in 2479 relatives. Blood 2009; 113; 5314-22.

73. Caspers M, Pavlova A, Driesen J, Harbrecht U, Klamroth R, Kadar J, Fischer R, Kemkes-Matthes B, Oldenburg J. Deficiencies of antithrombin, protein C and protein S - Practical experience in genetic analysis of a large patient cohort. Thromb Haemost 2012; 108; 247-57.

74. Luxembourg B, Delev D, Geisen C, Spannagl M, Krause M, Miesbach W, Heller C, Bergmann F, Schmeink U, Grossmann R, Lindhoff-Last E, Seifried E, Oldenburg J, Pavlova A. Molecular basis of antithrombin deficiency. Thromb Haemost 2011; 105; 635-46.

75. Aguila S, Martinez-Martinez I, Collado M, Llamas P, Anton AI, Martinez-Redondo C, Padilla J, Minano A, de la Morena-Barrio ME, Garcia-Avello A, Vicente V, Corral J. Compound heterozygosity involving Antithrombin

Bibliography

117

Cambridge II (p.Arg416Ser) in antithrombin deficiency. Thromb Haemost 2013; 109

76. Corral J, Hernandez-Espinosa D, Soria JM, Gonzalez-Conejero R, Ordonez A, Gonzalez-Porras JR, Perez-Ceballos E, Lecumberri R, Sanchez I, Roldan V, Mateo J, Minano A, Gonzalez M, Alberca I, Fontcuberta J, Vicente V. Antithrombin Cambridge II (A384S): an underestimated genetic risk factor for venous thrombosis. Blood 2007; 109; 4258-63.

77. Martinez-Martinez I, Navarro-Fernandez J, Ostergaard A, Gutierrez-Gallego R, Padilla J, Bohdan N, Minano A, Pascual C, Martinez C, de la Morena-Barrio ME, Aguila S, Pedersen S, Kristensen SR, Vicente V, Corral J. Amelioration of the severity of heparin-binding antithrombin mutations by posttranslational mosaicism. Blood 2012; 120; 900-4.

78. Hernandez-Espinosa D, Ordonez A, Vicente V, Corral J. Factors with conformational effects on haemostatic serpins: implications in thrombosis. Thromb Haemost 2007; 98; 557-63.

79. Corral J, Vicente V, Carrell RW. Thrombosis as a conformational disease. Haematologica 2005; 90; 238-46.

80. Yamasaki M, Li W, Johnson DJ, Huntington JA. Crystal structure of a stable dimer reveals the molecular basis of serpin polymerization. Nature 2008; 455; 1255-8.

81. Raja SM, Chhablani N, Swanson R, Thompson E, Laffan M, Lane DA, Olson ST. Deletion of P1 arginine in a novel antithrombin variant (antithrombin London) abolishes inhibitory activity but enhances heparin affinity and is associated with early onset thrombosis. J Biol Chem 2003; 278; 13688-95.

82. Martinez-Martinez I, Navarro-Fernandez J, Aguila S, Minano A, Bohdan N, de la Morena-Barrio ME, Ordonez A, Martinez C, Vicente V, Corral J. The infective polymerization of conformationally unstable antithrombin mutants may play a role in the clinical severity of antithrombin deficiency. Mol Med 2012; 18; 762-70.

83. Maclean PS, Tait RC. Hereditary and acquired antithrombin deficiency: epidemiology, pathogenesis and treatment options. Drugs 2007; 67; 1429-40.

84. Corral J, Huntington JA, Gonzalez-Conejero R, Mushunje A, Navarro M, Marco P, Vicente V, Carrell RW. Mutations in the shutter region of antithrombin result in formation of disulfide-linked dimers and severe venous thrombosis. J Thromb Haemost 2004; 2; 931-9.

85. Corral J, Aznar J, Gonzalez-Conejero R, Villa P, Minano A, Vaya A, Carrell RW, Huntington JA, Vicente V. Homozygous deficiency of heparin cofactor II: relevance of P17 glutamate residue in serpins, relationship with conformational diseases, and role in thrombosis. Circulation 2004; 110; 1303-7.

86. Abecasis GR, Cherny SS, Cookson WO, Cardon LR. Merlin--rapid analysis of dense genetic maps using sparse gene flow trees. Nat Genet 2002; 30; 97-101.

Bibliography

118

87. Corral J, Gonzalez-Conejero R, Soria JM, Gonzalez-Porras JR, Perez-Ceballos E, Lecumberri R, Roldan V, Souto JC, Minano A, Hernandez-Espinosa D, Alberca I, Fontcuberta J, Vicente V. A nonsense polymorphism in the protein Z-dependent protease inhibitor increases the risk for venous thrombosis. Blood 2006; 108; 177-83.

88. Sole X, Guino E, Valls J, Iniesta R, Moreno V. SNPStats: a web tool for the analysis of association studies. Bioinformatics 2006; 22; 1928-9.

89. Almasy L, Blangero J. Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet 1998; 62; 1198-211.

90. Conlan MG, Folsom AR, Finch A, Davis CE, Marcucci G, Sorlie P, Wu KK. Antithrombin III: associations with age, race, sex and cardiovascular disease risk factors. The Atherosclerosis Risk in Communities (ARIC) Study Investigators. Thromb Haemost 1994; 72; 551-6.

91. Llop E, Gallego RG, Belalcazar V, Gerwig GJ, Kamerling JP, Segura J, Pascual JA. Evaluation of protein N-glycosylation in 2-DE: Erythropoietin as a study case. Proteomics 2007; 7; 4278-91.

92. Mushunje A, Zhou A, Carrell RW, Huntington JA. Heparin-induced substrate behavior of antithrombin Cambridge II. Blood 2003; 102; 4028-34.

93. Hernandez-Espinosa D, Minano A, Martinez C, Perez-Ceballos E, Heras I, Fuster JL, Vicente V, Corral J. L-asparaginase-induced antithrombin type I deficiency: implications for conformational diseases. Am J Pathol 2006; 169; 142-53.

94. Mushunje A, Evans G, Brennan SO, Carrell RW, Zhou A. Latent antithrombin and its detection, formation and turnover in the circulation. J Thromb Haemost 2004; 2; 2170-7.

95. Helander A, Husa A, Jeppsson JO. Improved HPLC method for carbohydrate-deficient transferrin in serum. Clin Chem 2003; 49; 1881-90.

96. Cartegni L, Chew SL, Krainer AR. Listening to silence and understanding nonsense: exonic mutations that affect splicing. Nat Rev Genet 2002; 3; 285-98.

97. Fairbrother WG, Yeh RF, Sharp PA, Burge CB. Predictive identification of exonic splicing enhancers in human genes. Science 2002; 297; 1007-13.

98. Zhang XH, Chasin LA. Computational definition of sequence motifs governing constitutive exon splicing. Genes Dev 2004; 18; 1241-50.

99. Di MM, Thys C, Waelkens E, Overbergh L, D'Hertog W, Mathieu C, De VR, Peerlinck K, Van GC, Freson K. An integrated proteomics and genomics analysis to unravel a heterogeneous platelet secretion defect. J Proteomics 2011; 74; 902-13.

100. Di MM, Van GC, Freson K. Recent advances in platelet proteomics. Expert Rev Proteomics 2012; 9; 451-66.

Bibliography

119

101. Shevchenko A, Wilm M, Vorm O, Mann M. Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal Chem 1996; 68; 850-8.

102. Gharahdaghi F, Weinberg CR, Meagher DA, Imai BS, Mische SM. Mass spectrometric identification of proteins from silver-stained polyacrylamide gel: a method for the removal of silver ions to enhance sensitivity. Electrophoresis 1999; 20; 601-5.

103. Medina P, Navarro S, Estelles A, Vaya A, Woodhams B, Mira Y, Villa P, Migaud-Fressart M, Ferrando F, Aznar J, Bertina RM, Espana F. Contribution of polymorphisms in the endothelial protein C receptor gene to soluble endothelial protein C receptor and circulating activated protein C levels, and thrombotic risk. Thromb Haemost 2004; 91; 905-11.

104. Roldan V, Ordonez A, Marin F, Zorio E, Soria JM, Minano A, Espana F, Gonzalez-Conejero R, Pineda J, Estelles A, Fontcuberta J, Vicente V, Corral J. Antithrombin Cambridge II (A384S) supports a role for antithrombin deficiency in arterial thrombosis. Thromb Haemost 2009; 101; 483-6.

105. Severinsen MT, Johnsen SP, Tjonneland A, Overvad K, Dethlefsen C, Kristensen SR. Body height and sex-related differences in incidence of venous thromboembolism: a Danish follow-up study. Eur J Intern Med 2010; 21; 268-72.

106. Medina P, Navarro S, Corral J, Zorio E, Roldan V, Estelles A, Santamaria A, Marin F, Rueda J, Bertina RM, Espana F. Endothelial protein C receptor polymorphisms and risk of myocardial infarction. Haematologica 2008; 93; 1358-63.

107. Peyrard M, Seroussi E, Sandberg-Nordqvist AC, Xie YG, Han FY, Fransson I, Collins J, Dunham I, Kost-Alimova M, Imreh S, Dumanski JP. The human LARGE gene from 22q12.3-q13.1 is a new, distinct member of the glycosyltransferase gene family. Proc Natl Acad Sci U S A 1999; 96; 598-603.

108. McCoy AJ, Pei XY, Skinner R, Abrahams JP, Carrell RW. Structure of beta-antithrombin and the effect of glycosylation on antithrombin's heparin affinity and activity. J Mol Biol 2003; 326; 823-33.

109. de la Morena-Barrio ME, Anton AI, Martinez-Martinez I, Padilla J, Minano A, Navarro-Fernandez J, Aguila S, Lopez MF, Fontcuberta J, Vicente V, Corral J. Regulatory regions of SERPINC1 gene: Identification of the first mutation associated with antithrombin deficiency. Thromb Haemost 2012; 107; 430-7.

110. Jaeken J. Congenital disorders of glycosylation (CDG): it's (nearly) all in it! J Inherit Metab Dis 2011; 34; 853-8.

111. Grünewald S, Schollen E, Van Schaftingen E, Jaeken J, Matthijs G. High residual activity of PMM2 in patients' fibroblasts: possible pitfall in the diagnosis of CDG-Ia (phosphomannomutase deficiency). Am J Hum Genet 2001; 68; 347-54.

Bibliography

120

112. Matthijs G, Schollen E, Bjursell C, Erlandson A, Freeze H, Imtiaz F, Kjaergaard S, Martinsson T, Schwartz M, Seta N, Vuillaumier-Barrot S, Westphal V, Winchester B. Mutations in PMM2 that cause congenital disorders of glycosylation, type Ia (CDG-Ia). Hum Mutat 2000; 16; 386-94.

113. Savignac M, Edir A, Simon M, Hovnanian A. Darier disease : a disease model of impaired calcium homeostasis in the skin. Biochim Biophys Acta 2011; 1813; 1111-7.

114. Axelsson MA, Karlsson NG, Steel DM, Ouwendijk J, Nilsson T, Hansson GC. Neutralization of pH in the Golgi apparatus causes redistribution of glycosyltransferases and changes in the O-glycosylation of mucins. Glycobiology 2001; 11; 633-44.

115. Bjork I, Olson ST. Antithrombin. A bloody important serpin. Adv Exp Med Biol 1997; 425; 17-33.

116. Kottke-Marchant K, Duncan A. Antithrombin deficiency: issues in laboratory diagnosis. Arch Pathol Lab Med 2002; 126; 1326-36.

117. Austin H, De Staercke C., Lally C, Bezemer ID, Rosendaal FR, Hooper WC. New gene variants associated with venous thrombosis: a replication study in White and Black Americans. J Thromb Haemost 2011; 9; 489-95.

118. Bezemer ID, Bare LA, Doggen CJ, Arellano AR, Tong C, Rowland CM, Catanese J, Young BA, Reitsma PH, Devlin JJ, Rosendaal FR. Gene variants associated with deep vein thrombosis. JAMA 2008; 299; 1306-14.

119. Guerra-Shinohara EM, Bertinato JF, Tosin BC, Cordeiro da SK, Burlacchini de Carvalho MH, Pulcineli Vieira FR, Zugaib M, Cerda A, Morelli VM. Polymorphisms in antithrombin and in tissue factor pathway inhibitor genes are associated with recurrent pregnancy loss. Thromb Haemost 2012; 108; 693-700.

120. Tremp GL, Duchange N, Branellec D, Cereghini S, Tailleux A, Berthou L, Fievet C, Touchet N, Schombert B, Fruchart JC, . A 700-bp fragment of the human antithrombin III promoter is sufficient to confer high, tissue-specific expression on human apolipoprotein A-II in transgenic mice. Gene 1995; 156; 199-205.

121. Thomas DJ, Rosenbloom KR, Clawson H, Hinrichs AS, Trumbower H, Raney BJ, Karolchik D, Barber GP, Harte RA, Hillman-Jackson J, Kuhn RM, Rhead BL, Smith KE, Thakkapallayil A, Zweig AS, Haussler D, Kent WJ. The ENCODE Project at UC Santa Cruz. Nucleic Acids Res 2007; 35; D663-D667.

122. Tait RC, Walker ID, Islam SI, McCall F, Conkie JA, Mitchell R, Davidson JF. Influence of demographic factors on antithrombin III activity in a healthy population. Br J Haematol 1993; 84; 476-80.

123. Westrick RJ, Ginsburg D. Modifier genes for disorders of thrombosis and hemostasis. J Thromb Haemost 2009; 7 Suppl 1; 132-5.

Bibliography

121

124. Oudot-Mellakh T, Cohen W, Germain M, Saut N, Kallel C, Zelenika D, Lathrop M, Tregouet DA, Morange PE. Genome wide association study for plasma levels of natural anticoagulant inhibitors and protein C anticoagulant pathway: the MARTHA project. Br J Haematol 2012; 157; 230-9.

125. Kanagawa M, Saito F, Kunz S, Yoshida-Moriguchi T, Barresi R, Kobayashi YM, Muschler J, Dumanski JP, Michele DE, Oldstone MB, Campbell KP. Molecular recognition by LARGE is essential for expression of functional dystroglycan. Cell 2004; 117; 953-64.

126. Aguilan JT, Sundaram S, Nieves E, Stanley P. Mutational and functional analysis of Large in a novel CHO glycosylation mutant. Glycobiology 2009; 19; 971-86.

127. Patnaik SK, Stanley P. Mouse large can modify complex N- and mucin O-glycans on alpha-dystroglycan to induce laminin binding. J Biol Chem 2005; 280; 20851-9.

128. Brockington M, Torelli S, Prandini P, Boito C, Dolatshad NF, Longman C, Brown SC, Muntoni F. Localization and functional analysis of the LARGE family of glycosyltransferases: significance for muscular dystrophy. Hum Mol Genet 2005; 14; 657-65.

129. Hu Y, Li ZF, Wu X, Lu Q. Large induces functional glycans in an O-mannosylation dependent manner and targets GlcNAc terminals on alpha-dystroglycan. PLoS One 2011; 6; e16866.

130. Zhang P, Hu H. Differential glycosylation of alpha-dystroglycan and proteins other than alpha-dystroglycan by like-glycosyltransferase. Glycobiology 2012; 22; 235-47.

131. Inamori KI, Hara Y, Willer T, Anderson ME, Zhu Z, Yoshida-Moriguchi T, Campbell KP. Xylosyl- and glucuronyltransferase functions of LARGE in alpha-dystroglycan modification are conserved in LARGE2. Glycobiology 2012;

132. Aebi M, Bernasconi R, Clerc S, Molinari M. N-glycan structures: recognition and processing in the ER. Trends Biochem Sci 2010; 35; 74-82.

133. Smith JW, Knauer DJ. A heparin binding site in antithrombin III. Identification, purification, and amino acid sequence. J Biol Chem 1987; 262; 11964-72.

134. Jaeken J, Vanderschueren-Lodewyckx M, Snoeck L, Corbeel L, Wggermont E, Eeckels R. Familial psychomotor retardation with markedly fluctuating serum prolactin, FSH and GH levels, partial TBG deficiency, increased serum arylsulphatase A and increased CSF protein: a new syndrome). Pediatr Res 1980; 14; 179.

135. Jaeken J, Matthijs G. Congenital disorders of glycosylation. Annu Rev Genomics Hum Genet 2001; 2; 129-51.

136. Jaeken J. Congenital disorders of glycosylation. Ann N Y Acad Sci 2010; 1214; 190-8.

Bibliography

122

137. Arnoux JB, Boddaert N, Valayannopoulos V, Romano S, Bahi-Buisson N, Desguerre I, de Keyzer Y., Munnich A, Brunelle F, Seta N, Dautzenberg MD, de Lonlay P. Risk assessment of acute vascular events in congenital disorder of glycosylation type Ia. Mol Genet Metab 2008; 93; 444-9.

138. Young G, Driscoll MC. Coagulation abnormalities in the carbohydrate-deficient glycoprotein syndrome: case report and review of the literature. Am J Hematol 1999; 60; 66-9.

139. Van Geet C, Jaeken J, Freson K, Lenaerts T, Arnout J, Vermylen J, Hoylaerts MF. Congenital disorders of glycosylation type Ia and IIa are associated with different primary haemostatic complications. J Inherit Metab Dis 2001; 24; 477-92.

140. Jaeken J, Hennet T, Matthijs G, Freeze HH. CDG nomenclature: time for a change! Biochim Biophys Acta 2009; 1792; 825-6.

141. Aebi M, Helenius A, Schenk B, Barone R, Fiumara A, Berger EG, Hennet T, Imbach T, Stutz A, Bjursell C, Uller A, Wahlstrom JG, Briones P, Cardo E, Clayton P, Winchester B, Cormier-Dalre V, de LP, Cuer M, Dupre T, et al. Carbohydrate-deficient glycoprotein syndromes become congenital disorders of glycosylation: an updated nomenclature for CDG. First International Workshop on CDGS. Glycoconj J 1999; 16; 669-71.

142. Eklund EA, Freeze HH. The congenital disorders of glycosylation: a multifaceted group of syndromes. NeuroRx 2006; 3; 254-63.

143. Burrows NP, Nicholls AC, Richards AJ, Luccarini C, Harrison JB, Yates JR, Pope FM. A point mutation in an intronic branch site results in aberrant splicing of COL5A1 and in Ehlers-Danlos syndrome type II in two British families. Am J Hum Genet 1998; 63; 390-8.

144. Khan SG, Metin A, Gozukara E, Inui H, Shahlavi T, Muniz-Medina V, Baker CC, Ueda T, Aiken JR, Schneider TD, Kraemer KH. Two essential splice lariat branchpoint sequences in one intron in a xeroderma pigmentosum DNA repair gene: mutations result in reduced XPC mRNA levels that correlate with cancer risk. Hum Mol Genet 2004; 13; 343-52.

145. Vuillaumier-Barrot S, le BC, de LP, Madinier-Chappat N, Barnier A, Dupre T, Durand G, Seta N. PMM2 intronic branch-site mutations in CDG-Ia. Mol Genet Metab 2006; 87; 337-40.

146. Jones MA, Ng BG, Bhide S, Chin E, Rhodenizer D, He P, Losfeld ME, He M, Raymond K, Berry G, Freeze HH, Hegde MR. DDOST mutations identified by whole-exome sequencing are implicated in congenital disorders of glycosylation. Am J Hum Genet 2012; 90; 363-8.

147. Landberg E, Pahlsson P, Lundblad A, Arnetorp A, Jeppsson JO. Carbohydrate composition of serum transferrin isoforms from patients with high alcohol consumption. Biochem Biophys Res Commun 1995; 210; 267-74.

Bibliography

123

148. Cottalasso D, Bellocchio A, Pronzato MA, Domenicotti C, Traverso N, Gianelli MV, Marinari UM, Nanni G. Effect of ethanol administration on the level of dolichol in rat liver microsomes and Golgi apparatus. Alcohol Clin Exp Res 1998; 22; 730-7.

149. Flahaut C, Michalski JC, Danel T, Humbert MH, Klein A. The effects of ethanol on the glycosylation of human transferrin. Glycobiology 2003; 13; 191-8.

150. Landberg E, Pahlsson P, Lundblad A, Arnetorp A, Jeppsson JO. Carbohydrate composition of serum transferrin isoforms from patients with high alcohol consumption. Biochem Biophys Res Commun 1995; 210; 267-74.

151. Guizzetti M, Chen J, Oram JF, Tsuji R, Dao K, Moller T, Costa LG. Ethanol induces cholesterol efflux and up-regulates ATP-binding cassette cholesterol transporters in fetal astrocytes. J Biol Chem 2007; 282; 18740-9.

152. Krebs HA, Freedland RA, Hems R, Stubbs M. Inhibition of hepatic gluconeogenesis by ethanol. Biochem J 1969; 112; 117-24.

153. Sharma V, Ichikawa M, He P, Scott DA, Bravo Y, Dahl R, Ng BG, Cosford ND, Freeze HH. Phosphomannose isomerase inhibitors improve N-glycosylation in selected phosphomannomutase-deficient fibroblasts. J Biol Chem 2011; 286; 39431-8.

Spanish abstract/Resumen en castellano

125

Spanish abstract

Resumen en castellano I


126


127

La importancia del sistema hemostático en la enfermedad tromboembólica, la

principal causa de morbimortalidad de nuestra sociedad, ha incentivado la búsqueda de

alteraciones genéticas asociadas con variaciones en el funcionamiento del sistema

hemostático, empleando diferentes diseños experimentales. Gran parte de estos estudios

se han centrado en los propios elementos de dicho sistema, con resultados bastante

frustrantes. La elevada variabilidad interindividual de los fenotipos hemostáticos, junto

con la alta heredabilidad de los mismos, sugieren la posible existencia de variabilidad

genética en el gen que codifica dicho fenotipo. También alteraciones genéticas en otros

factores podrían modular de forma indirecta los niveles de ese elemento. En cualquier

caso, estas alteraciones pueden jugar un papel relevante en la funcionalidad del sistema

hemostático y en el riesgo trombótico.

La antitrombina es un miembro de la superfamilia de las serpinas, inhibidores de

serín proteasas, clave en el control de la hemostasia. Su amplio espectro de proteasas

procoagulantes diana, que abarca moléculas tan relevantes como el FXa o la trombina,

pero también otras como el FVII, FIXa, FXIa y FXIIa entre otras, y el mecanismo de

inhibición tan potente y eficaz, hacen que la antitrombina sea el principal anticoagulante

endógeno del organismo. Por ello, cualquier factor que afecte a esta molécula es

susceptible de desequilibrar el sistema hemostático, y podría contribuir a incrementar el

riesgo trombótico. La búsqueda de estos elementos se ha centrado clásicamente en el

análisis molecular del gen codificante: SERPINC1. Así desde que en 1965 O Egeberg

describió la primera familia con deficiencia de antitrombina asociada a una alta

incidencia de trombosis venosa, se han identificado más de 296 mutaciones diferentes

en el gen SERPINC1 en pacientes con trombosis venosa y deficiencia de antitrombina.

Todas estas mutaciones afectan la región codificante o secuencias de procesamiento de

intrones y explican, por distintos mecanismos, los tipos de deficiencia de antitrombina.

Sin embargo, es sorprendente la ausencia de modificaciones genéticas que a través de

mecanismos regulatorios se asocian con deficiencia de antitrombina. Probablemente las

mutaciones que afecten a estas regiones son poco frecuentes o tienen poca relevancia

funcional. Por otra parte, no se conocen las regiones promotora y/o reguladoras del gen

SERPINC1. Hasta la fecha, solo un estudio de nuestro grupo ha descrito un

polimorfismo localizado en una zona no codificante, en el intrón 1, que regula

moderadamente los niveles de antitrombina.

Por el contrario, apenas hay estudios que hayan buscado otros elementos fuera del

gen codificante, capaces de afectar los niveles o funcionalidad de este relevante


128

anticoagulante. Este dato junto al reciente trabajo que demuestra que la reducción

incluso moderada de los niveles de antitrombina también incrementa el riesgo

trombótico, anima a buscar nuevos mecanismos que potencialmente afecten la

funcionalidad o los niveles de antitrombina.

El objetivo principal de este estudio fue identificar nuevos genes y modificaciones

genéticas involucradas en la regulación de los niveles o funcionalidad de antitrombina.

Para llevar a cabo este objetivo utilizamos tres aproximaciones diferentes.

1. Identificación de nuevas regiones reguladoras del gen SERPINC1 mediante la

secuenciación completa del gen (15,000 pb, abarcando todos los exones e intrones de

este gen, así como las regiones 3’ y 5’) en tres grupos de sujetos con diferentes niveles

de antitrombina.

2. Identificación de nuevos genes moduladores de antitrombina empleando un

estudio de asociación de polimorfismos que abarcan el genoma completo (Genome

Wide Association Study, GWAS) y que se complementó con diferentes estudios

experimentales, incluyendo silenciamiento del gen identificado.

3. Identificación de nuevos mecanismos responsables de la deficiencia de

antitrombina, mediante la caracterización bioquímica de la antitrombina plasmática de

pacientes con deficiencia de antitrombina.

Para el desarrollo del proyecto hemos empleado metodología molecular que

abarca desde genotipado de alteraciones puntuales empleando diferentes sistemas

(sondas taqman, PCR-ASRA, etc), hasta un GWAS, pasando por la cuantificación de

expresión y la secuenciación de diferentes genes (SERPINC1, LARGE, PMM2).

También hemos usado metodología bioquímica clásica (electroforesis en geles

desnaturalizantes y nativos, purificación o separación de proteínas mediante FPLC o

HPLC, estudios de inhibición, etc), pero también metodología más avanzada

(proteómica y glicómica). Finalmente, en este proyecto también hemos aplicado

metodología celular que incluye inmunofluorescencia, estudios de reporter, o

silenciamiento génico.

Desde el punto de vista metodológico también debemos destacar el amplio

espectro de pacientes y controles estudiados. Así, el proyecto ha incluido una cohorte

familiar (estudio GAIT), 307 donantes de sangre sanos, tres estudios caso-control de


129

diferentes enfermedades tromboembólicas (infarto agudo de miocardio y trombosis

venosa) de dos países diferentes (España y Dinamarca) con 2,980 pacientes y 3,996

controles. Destacamos la inclusión de 118 pacientes con deficiencia de antitrombina,

grupo del que seleccionamos 25 pacientes que con deficiencia (moderada o severa) de

antitrombina no tenían mutaciones en el gen SERPINC1 y han sido la base de dos de los

apartados de esta tesis.

Finalmente estudiamos también 2 pacientes con trastorno de glicosilación

congénita (CDG), un desorden muy raro y un paciente con severo alcoholismo,

cruciales para el último apartado de nuestro estudio.

Los resultados de este estudio muestran una heterogeneidad de nuevos

mecanismos implicados en la regulación de los niveles de antitrombina con potenciales

consecuencias patológicas.

1) El análisis del gen SERPINC1 completo en sujetos con deficiencia de

antitrombina pero sin mutaciones en la zona codificante de dicho gen identificó la

primera mutación en la región promotora del gen asociada con deficiencia de

antitrombina. El cambio g.2143 C>G identificado en la familia 2 no impide la

transcripción del gen pero sí la reduce de forma significativa, como muestran los niveles

de antitrombina en portadores y los estudios con reporter de luciferasa en diferentes

líneas celulares. El cambio g.2143 C>G se encuentra en una de las cuatro regiones

protegidas descritas en dos estudios independientes; y se ha sugerido que en esta región

podría interaccionar el factor nuclear hepático 4 (HNF-4) regulando la transcripción del

gen. Además, en un paciente con moderada deficiencia de antitrombina identificamos

otra nueva mutación g.1091 C>T situada a 1,052 pb del codón ATG, con potencial

efecto regulatorio. Son necesarios estudios familiares y de reporter de luciferasa para

determinar la relevancia funcional de esta segunda alteración en la regulación del gen.

Estos resultados sugieren incluir el estudio de la región promotora del gen

SERPINC1 en la caracterización molecular de la deficiencia de antitrombina,

concretamente en aquéllos casos con deficiencia tipo I y sin mutaciones en las regiones

codificantes de SERPINC1.

Sin embargo, es probable la existencia de otros elementos reguladores aguas

arriba o abajo del gen SERPINC1. La identificación de una familia (F1) y un paciente

(P26) pertenecientes a la misma área geográfica, con deficiencia tipo I y trombosis


130

severas y sin mutaciones en las 15,925 pb analizadas de SERPINC1, alienta a continuar

la búsqueda de nuevas regiones reguladoras.

La búsqueda de esas regiones reguladoras también la hemos realizado con otro

diseño experimental, centrándonos en población general ya que los niveles de este

anticoagulante tienen una alta variabilidad y existe un elevado grado de heredabilidad

para este fenotipo (h=0.486) lo que sustenta la existencia de factores genéticos

implicados. Estudiamos los niveles de antitrombina en una muestra de 307 sujetos

sanos. Como se esperaba encontramos una distribución normal (70-130%).

Seleccionamos cinco sujetos con niveles extremos de antitrombina dentro del rango de

la normalidad (75% y 115%) y secuenciamos el gen SERPINC1 completo con el

objetivo de identificar variaciones genéticas que se asocien con esa gran variabilidad.

Las alteraciones identificadas con potencial efecto funcional fueron verificadas en una

muestra mayor. Los resultados confirmaron el moderado efecto que el polimorfismo

rs2227589 del intrón 1 tiene sobre los niveles de antitrombina. También identificamos

un polimorfismo de la región promotora g.2085 T>C que in silico modificaba

potencialmente la unión de distintos factores de transcripción. Sin embargo, el estudio

de asociación de actividad y frecuencia de este cambio en 131 sujetos sanos demostró

no estar asociado a los niveles de antitrombina.

La disección de la arquitectura del gen SERPINC1 en sujetos sanos con niveles

extremos de antitrombina reveló dos importantes hallazgos: A) La variabilidad genética

de este gen es muy baja. Únicamente encontramos 22 polimorfismos, y solo 2 en la

región codificante, ambos ligados y sin que modifiquen la secuencia aminoacídica.

Estos resultados reflejan la sensibilidad conformacional y funcional de esta molécula.

B) La escasa importancia de esta variabilidad genética en el control de los niveles de

antitrombina.

Estos resultados junto al alto grado de heredabilidad de los niveles de

antitrombina apoyan la existencia de genes moduladores. Estos genes podrían codificar

factores transcripcionales, moléculas implicadas en el plegamiento o secreción de

antitrombina, enzimas implicadas en modificaciones post-traduccionales, proteasas,

miRNAs…

2) Por ello, nuestro segundo objetivo fue buscar nuevos genes moduladores de

antitrombina. Para ello empleamos una aproximación de asociación genotipo-fenotipo

analizando el genoma completo (GWAS). Esta misma aproximación se acaba de

publicar recientemente por un grupo francés. Estos estudios se diferenciaban en la


131

población estudiada: pacientes con trombosis (estudio francés) o nuestro estudio

familiar. Pero en ambos estudios, el GWAS no identificó ningún polimorfismo asociado

con los niveles de antitrombina con significación estadística (p< 10E-07). Este

resultado, junto a la elevada heredabilidad, sugería la existencia de elementos genético

con poco peso, por lo que nosotros validamos los 10 SNPs con mayor asociación

estadística (p< 4x10E-05) y dos SNPs adicionales del gen LARGE, un gen que codifica

una glicosiltransferasa que se encontraba en la lista de los 10 genes mas asociados con

los niveles de antitrombina. Precisamente un SNP de LARGE, rs762057, mantuvo la

asociación con los niveles de anti-FXa en plasma en los 307 donantes de sangre de otra

región española que se emplearon como población de validación. Tanto en el estudio de

validación como en el GWAS, el alelo polimórfico incrementaba moderadamente los

niveles de antitrombina (2-4%). Este ligero aumento de la actividad de antitrombina

explica la ausencia de relevancia trombótica de este polimorfismo, según reflejaron los

resultados del estudio de 849 pacientes con trombosis venosa y 862 controles. Además,

encontramos que el haplotipo H2 del gen LARGE definido con los tres polimorfismos

rs762057, rs713703 y rs240082 también se asociaba de forma significativa con los

niveles de antitrombina (p=0.030). Para consolidar esta asociación y definir el

mecanismo de la misma, realizamos experimentos de silenciamiento del gen LARGE en

dos líneas celulares (HepG2, línea hepática y con expresión constitutiva de antitrombina

y HEK-EBNA transfectada con cDNA de antitrombina humana). El silenciamiento de

LARGE redujo la secreción de antitrombina en ambas líneas celulares confirmando el

papel de LARGE en la modulación de los niveles de antitrombina. El mecanismo por el

que actúa LARGE, una enzima con actividades xilosiltransferasa y

glucuroniltransferasa, sobre los niveles de antitrombina tiene todavía que explorarse,

pero el hecho de que esta enzima genere moléculas parecidas a heparina y que la

antitrombina presente un domino de unión a heparina, sustenta esta asociación y

sugieren que LARGE podría estar implicado en el correcto plegamiento o tráfico

intracelular de la antitrombina.

3) La caracterización bioquímica de la antitrombina plasmática en pacientes con

deficiencia de antitrombina sin mutaciones en SERPINC1 ha sido crucial para

identificar un nuevo mecanismo implicado en la regulación de los niveles de

antitrombina y con gran riesgo trombótico. Este particular factor trombofílico afecta una

modificación postraduccional, la N-glicosilación. El paciente P27 tuvo cuatro episodios


132

trombóticos en edades tempranas y una significativa deficiencia de antitrombina como

único factor trombofílico. La identificación de diferentes glicoformas de antitrombina

mediante análisis electroforético seguido de diferentes estudios bioquímicos y un

profundo análisis de secuenciación del gen PMM2, gen que codifica la enzima

fosfomanomutasa 2, implicada en los primeros pasos de formación del N-glicano

precursor que se incorporará a N-glicoproteínas nacientes, permitió el diagnóstico final

de un trastorno de glicosilación congénita tipo Ia (CDGIa o también llamado PMM2-

CDG). Además, este nuevo factor trombofílico, más bien una variante, podría ser

relativamente frecuente. El análisis electroforético de antitrombina y antitripsina y el

estudio de glicoformas de transferrina mediante HPLC y Q-TOF en 25 pacientes con

trombosis y ligera deficiencia de antitrombina sin mutaciones en SERPINC1, identificó

cuatro pacientes con un patrón igual pero más suave que el de CDGs. En tres casos se

identificó una mutación en heterocigosis en PMM2. Estos resultados sugieren la

existencia de un nuevo desorden al que hemos llamado CDG-like. Como los cuatro

pacientes tienen en común una leve afectación hepática como consecuencia de la ingesta

de alcohol, especulamos que esta combinación podría explicar el fenotipo CDG y la

clínica trombótica. En su conjunto estos resultados apoyan que los trastornos de

glicosilación pudieran considerarse como factor trombofílico que debería estudiarse,

especialmente en pacientes con moderada deficiencia de antitrombina.

Nuestro estudio muestra nuevo mecanismos implicados en la variabilidad de la

antitrombina, el principal anticoagulante endógeno. Además, abre nuevas líneas de

trabajo y deja cuestiones que aún quedan por resolver: la existencia de nuevas regiones

reguladoras en el gen SERPINC1 que aún deben ser identificadas, la existencia de otros

genes moduladores de antitrombina que podrían tener relevancia trombótica, la

incidencia de trastornos de glicosilación en trombosis, y el mecanismo por el cual estos

trastornos (PMM2-CDG y CDG-like) elevan el riesgo trombótico.

Appendix I

133

Diffusion of scientific results related with the Thesis

Appendix I

Appendix I

134

Appendix I

135

Results from the present Thesis have been published or are under review in peer-

reviewed journals.

A. Published:

1. de la Morena-Barrio ME, Anton AI, Martinez-Martinez I, Padilla, J, Minano

A, Navarro-Fernández J, Águila S, López MF, Fontcuberta J, Vicente V, and Corral J.

Regulatory regions of SERPINC1 gene: identification of the first mutation associated

with antithrombin deficiency. Thromb. Haemost. 107(3):430-7; 2012. Impact Factor:

5.044.

2. de la Morena-Barrio ME, Sevivas T, Martínez-Martínez I, Miñano A, Vicente

V, Jaeken J, Corral J. Congenital disorder of glycosylation (PMM2-CDG) in a patient

with antithrombin deficiency and severe thrombophilia. J Thromb Haemost. 2012 Oct

20. doi: 10.1111/jth.12031. Impact factor: 5.731.

B. Submitted:

1. de la Morena-Barrio ME, Buil A, Antón AI, Martínez-Martínez I, Miñano A,

Gutiérrez-Gallego R, Navarro-Fernández J, Águila S, Souto JC, Vicente V, Soria JM,

Corral J. Identification of antithrombin-modulating genes. Role of LARGE, a gene

encoding a bifunctional glycosyltransferase, in the secretion of proteins?. PLoS ONE.

Impact factor: 4.092

Appendix I

136

137

Blood Coagulation, Fibrinolysis and Cellular Haemostasis © Schattauer 2012 1

Regulatory regions of SERPINC1 gene: Identification of the first mutation associated with antithrombin deficiency

María Eugenia de la Morena-Barrio1; Ana Isabel Antón1; Irene Martínez-Martínez1; José Padilla1; Antonia Miñano1; José Navarro-Fernández1; Sonia Águila1; María Fernanda López2; Jordi Fontcuberta3; Vicente Vicente1; Javier Corral1 1Centro 3Unitat

Regional de Hemodonación, University of Murcia, Spain; 2Unidad de Hemostasia y Trombosis del Complejo Hospitalario Universitario, A Coruña, Spain; d'Hemostàsia i Trombosi, Departament d'Hematologia, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain

Summary Antithrombin is the main endogenous anticoagulant. Impaired function or deficiency of this molecule significantly increases the risk of thrombosis. We studied the genetic variability of SERPINC1, the gene encoding antithrombin, to identify mutations affecting regulatory regions with functional effect on its levels. We sequenced 15,375 bp of this gene, including the potential promoter region, in three groups of subjects: five healthy subjects with antithrombin levels in the lowest (75%) and highest (115%) ranges of our population, 14 patients with venous thrombosis and a moderate antithrombin deficiency as the single thrombophilic defect, and two families with type I antithrombin deficiency who had neither mutations affecting exons or flanking regions, nor gross gene deletions. Our study confirmed the low genetic variability of SERPINC1, particularly in the coding region, and its minor influence in the heterogeneity of antithrombin levels. Interestingly, in one

family, we identified a g.2143 C>G transversion, located 170 bp upstream from the translation initiation codon. This mutation affected one of the four regions located in the minimal promoter that have potential regulatory activity according to previous DNase footprinting protection assays. Genotype-phenotype analysis in the affected family and reporter analysis in different hepatic cell lines demonstrated that this mutation significantly impaired, although it did not abolish, the downstream transcription. Therefore, this is the first mutation affecting a regulatory region of the SERPINC1 gene associated with antithrombin deficiency. Our results strongly sustain the inclusion of the promoter region of SERPINC1 in the molecular analysis of patients with antithrombin deficiency.

Keywords SERPINC1, antithrombin, promoter, polymorphisms, thrombosis

Correspondence to: Dr. Javier Corral University of Murcia Centro Regional de Hemodonación Ronda de Garay S/N. Murcia 30003, Spain Tel.: +34 968341990, Fax: +34 968261914 E-mail: [email protected]

Financial support: This work was supported by 04515 /GERM/ 06 (Fundación Séneca), SAF2009–08993 (MCYT & FEDER), Fundación Mutua Madrileña, and RECAVA RD06/0014/0039 & RD06/0014/0016 (ISCIII & FEDER). MEMB is a holder of a predoctoral research grant from ISCIII (FI09/00190). IMM is a researcher from Fundación para la Formación e Investigación Sanitarias. JNF is postdoctoral researcher of the University of Murcia.

Received: October 11, 2011 Accepted after minor revision: November 29, 2011 Prepublished online: January 11, 2012 doi:10.1160/TH11-10-0701 Thromb Haemost 2012; 107: ■■■

Introduction

Antithrombin is an anticoagulant serpin essential for the haemostatic system. This plasma protein synthesised in the liver has an efficient mechanism of inhibition of multiple procoagulant proteases, including thrombin and factor (F) Xa (1, 2). Accordingly, complete deficiency of antithrombin has lethal consequences and the heterozygous deficiency significantly increases (10– to 40-fold) the risk of thrombosis (3). Thus, the screening of antithrombin deficiency evaluating the activity of antithrombin in plasma is included among the thrombophilic tests (4). SERPINC1, the gene encoding antithrombin (GenBank X68793.1) was mapped on chromosome 1q23–25.1 and comprises seven exons, encompassing 13.5 kb of genomic DNA (5, 6). More than 235 different gross gene deletions and mutations affecting the coding or splicing regions of SERPINC1 have been identified in pa-

tients with antithrombin deficiency (http://www.hgmd.cf.ac.uk/ac/gene.php?gene=SERPINC1). Different mechanisms, from RNA instability to conformational or functional consequences have been proposed to explain the associated quantitative or qualitative deficiency (3, 7). However, no mutation has been so far identified in regulatory regions. Actually, the mechanisms underlying antithrombin gene expression are not well known, the promoter lacks of TATA element or GC-rich regions (6, 8, 9) and few studies, most of them using deletional analysis and reporter or DNase I footprint assays, have identified few regions potentially involved in the transcriptional regulation of this gene (9–12). The aim of this study was the identification of gene variations affecting regulatory elements of SERPINC1 by sequencing this gene in three groups of subjects with different expression of antithrombin.

Thrombosis and Haemostasis 107.3/2012

Downloaded from www.thrombosis-online.com on 2012-02-14 | ID: 1000465936 | IP: 134.58.253.55

138

2 de la Morena-Barrio et al. First mutation in the promoter gene of antithrombin


Subjects and blood sampling

Our study included three groups of subjects: ● Three hundred seven Spanish Caucasian healthy blood donors (138 male/169 female), with an average age of 43 years. ● Fourteen thrombophilic patients (named as P1-P14) who had a moderate antithrombin deficiency (70–90% of the reference value) as the single thrombophilic risk factor identified. ● Two families (F1 and F2) with at least three members from two generations that repetitively had type I antithrombin deficiency (Fig. 2). These families were selected from 64 consecutive subjects with congenital antithrombin deficiency referenced to our laboratory from different Spanish and Portuguese hospitals. Selection was based on the absence of gross deletion, mutation in the coding region, or splicing site mutations.

Blood samples were obtained by venopuncture collection into 1:10 volume of trisodium citrate. Platelet-poor plasma fractions were obtained (within 5 minutes [min] after blood collection) by centrifugation at 4 ºC for 20 min at 2,200 x g and stored at –70ºC. Genomic DNA was purified by the salting out procedure and stored at –20ºC. All subjects gave their informed consent to enter the study, which was approved by the ethics committee of the Centro Regional de Hemodonación (Murcia, Spain) and performed according to the declaration of Helsinki, as amended in Edinburgh in 2000.

tems, Madrid, Spain). Comparison with the reference sequence (GenBank accession number NG_012462.1) was performed with SeqScape v2.5 software (Applied Biosystems).

Genotyping methods

The genotype of rs5778785 was determined by high resolution melting (HRM) analysis following the Type-it HRM PCR protocol (Qiagen, Madrid, Spain). The g.2085 T>C transition was genotyped by a HRM/razor probe technique. The genotype was verified by sequencing, with complete agreement. These two HRM analyses were carried out in Rotor-Gene Q 6000 (Qiagen) with the primers and probes shown in Suppl. Table 2 (available online at www.thrombosis-online.com). Nine polymorphisms were genotyped by PCR-allelic specific restriction assays using primers and restriction enzymes described in Table 1, as previously described (14). For some polymorphisms, mutated primers were designed (see Suppl. Table 2, available online at www.thrombosis-online.com). rs2227589 was genotyped by a Taqman SNP Genotyping Assay (C__16180170_20, Applied-Biosystem), as previously described (16). The g.2143 C>G mutation was studied by PCR-allelic specific restriction assay using: 5’-GGAGTCCTTGATCACACAGCA-3’ forward and 5’-GTCTTTGACTGTAACTACCAG-3’ reverse primers and Bsa JI restriction enzyme (New England BioLabs, Barcelona, Spain) following manufacturer’s specifications. Results were confirmed by sequencing.

Antithrombin levels and other thrombophilic tests

Prediction of transcription factor binding sites

Routine thrombophilic tests including protein C, protein S, anti phospholipid antibodies, FV Leiden, and prothrombin G20210A were performed, as described elsewhere (13). FXa-inhibiting activity from plasma of patients and family members was measured using a chromogenic method (HemosIL TH, Instrumentation Laboratory, Milan, Italy). Antithrombin antigen levels in plasma were determined by rocket immunoelectrophoresis. For both parameters, values were expressed as a percentage of the result observed in a pool of citrated plasma from 100 healthy subjects.

The search for sequences with potential transcriptional relevance affected by genetic variations identified in this study was performed by TESS and TFSEARCH programs, based on TRANSFAC databases (http://www.cbil.upenn.edu/cgi-bin/tess/tessand http://www.cbrc.jp/research/db/TFSEARCH).

Cloning and reporter assay

The 5’ non transcribed portion of SERPINC1 from one individual heterozygous for the g.2143 C>G mutation was amplified using Expand High Fidelity DNA polymerase (Roche) to minimise nucleotides misincorporation with 5’-GGAGTCCTTGATCACACAGCA-3’ (forward) and 5’-AATCTCGCAGAGGTTCCAGA-3’ (reverse) primers. PCR products were ligated into a T/A cloning vector (pCR®2.1, Invitrogen, Barcelona, Spain). Clones with wildtype and mutant alleles were selected by PCR-allelic restriction assay with Bsa JI, and the inserts cloned into Kpn I and Sac I sites of luciferase pGL3 basic vector (Promega, Madrid, Spain). Compet-

© Schattauer 2012

Amplification and sequencing Polymerase chain reactions (PCR) covering the whole SERPINC1 gene and the potential promoter region were carried out using Expand Long Template Polymerase (Roche, Madrid, Spain) and the oligonucleotide sets described in Suppl. Table 1 (available online at www.thrombosis-online.com). PCR products were purified and sequenced with ABI Prism Big Dye Terminator v3.1 Cycle sequencing kit and resolved on a 3130xl Genetic Analyzer (Applied Biosys-



139

de la Morena-Barrio et al. First mutation in the promoter gene of antithrombin 3

Table 1: Plasma anti- FXa values of blood donors according to the genotype of SER- PINC1 polymorphisms identified in five selected healthy subjects with extreme antithrombin levels and 14 patients with low antithrombin levels.

Polymorphism Subject Genotyping method

HRM/Razor probe & sequencing PCR-ASRA: Mnl I

Taqman probe

PCR-ASRA: Fok I

PCR-ASRA: Tsp 45I

PCR-ASRA: Fok I

PCR-ASRA: Hinf I

PCR-ASRA: Nla III

HRM & sequencing

PCR-ASRA: Fau I

PCR-ASRA: Dde I

PCR-ASRA: Bst EII

N* Frequency

PGA European# Our study

Anti-FXa activity (mean ± SD) WT:96 ± 9 Pol:99 ± 11 WT:95 ± 8 Pol:93 ± 9 WT:97 ± 7 Pol:95 ± 8 WT:95 ± 8 Pol:94 ± 9 WT:94 ± 8 Pol:97 ± 9 WT:94 ± 8 Pol:100 ± 19 WT:97 ± 7 Pol:95 ± 8 WT:95 ± 8 Pol:95 ± 9 WT:95 ± 8 Pol:95 ± 9 WT:95 ± 8 Pol:94 ± 8 WT:95 ± 8 Pol:94 ± 9 WT:94 ± 7 Pol:94 ± 8

P°

g.2085 T>C (rs78972925) rs2227588

rs2227589

rs2227590

rs2227596

rs2227603

rs941988

rs2227607

rs5778785

rs2759328

rs677

rs2227616

H1

P1 & P7

L1

L2

H1

H1

L1

H1

H1

L1

L2

H1

131

126

305

128

127

101

97

138

106

149

158

183

No data

A:0.875 G:0.125 G:0.975 A:0.043 C:0.870 T:0.130 A:0.795 G:0.205 T:0.935 G:0.065 G:0.955 A:0.045 G:0.972 T:0.028 No data

G:0.917 A:0.083 G:0.842 C:0.158 A:0.935 (-):0.065

T:0.973 C:0.027 A:0.921 G:0.079 G:0.898 A:0.102 C:0.922 T:0.078 A:0.886 G:0.114 T:0.990 G:0.010 G:0.878 A:0.122 G:0.895 T:0.105 WT:0.896 INS:0.104 G:0.850 A:0.150 G:0.930 C:0.070 A:0.880 (-):0.120

0.375

0.342

0.037

0.793

0.129

0.317

0.513

0.903

0.838

0.736

0.604

0.923

* N: Number of healthy blood donors studied. # Frequencies from http://www.ncbi.nlm.nih.gov/snp.° p-value obtained from the student-t statistical analysis following a dominant model. WT: Homozygous wild type genotype; Pol: Carriers of the polymorphism (heterozygous + homozygous). SD: standard deviation. rs941988 and rs2227592 were completely linked, according to the Haplo- view 4.1. The genotypes of all studied polymorphisms were in Hardy-Weinberg equilibrium.

ent cells were transformed, plasmid DNA was purified using a commercial kit (Plasmid Midi Kit, Qiagen), and they were sequenced in both directions to faithfully confirm the wild-type and mutant sequences. The ability of each sequence to promote transcription of luciferase gene was transiently tested in human cell lines PLC-PRF-5 (human hepatome cell line) and HepG2 (human hepatocellular liver carcinoma cell line) both with antithrombin constitutive expression. All plasmid DNA were quantified by Nanodrop (Thermo Scien- tific, Madrid, Spain). Eighteen hours before transfection, cells were plated at 80,000 per well of a 24-well plated (three replicates per clone for each cell line and in each experiment) and cultured in Dulbecco's Modified Eagle Medium (Invitrogen), at 37ºC with 5% CO2. Reporter plasmid (1,000 ng) along with 100 ng of Renilla luciferase control plasmid pRL-TK (Promega) were transfected by using Lipofectamine 2000 (Invitrogen) for PLC-PRF-5 cell line

© Schattauer 2012

and GenJet reagent (SignaGen® Laboratories) for HepG2 cell line, according to the manufacturer’s instructions. Luciferase assays were performed 48 h after transfection by using the Dual-GloTM Luciferase assay System (Promega). Promoter activity was normal- ised by dividing luciferase activity by Renilla luciferase activity for each transfected well. pGL3 control vector (Promega) was used as a transfection positive control.

Statistical analyses

Allele and genotype frequencies, deviations from Hardy-Weinberg expectations, haplotype analysis, association of haplotypes with anti-FXa activity and linkage disequilibrium analysis were calculated with the SNPstats software (15). Statistical analysis was performed with the SPSS software (ver-




sion 15.0). Data are presented as mean ± standard deviation. Dif- ferences in plasma anti-FXa activity between genotypes were analysed by means of Student's t-test using a dominant model.

Results

Genetic study in healthy population

The anti-FXa activity observed in the 307 Spanish Caucasian healthy blood donors had a normal distribution with a mean value of 97 ± 8% (see Suppl. Fig. 1, available online at www.thrombosis-online.com). None of the subjects had antithrombin deficiency (<70% of the reference value). We selected five subjects with extreme antithrombin levels within the normal range: three with the lowest levels (L1: 73.5%, L2: 74.5% and L3: 75.1%) and two with highest levels (H1: 114.4% and H2: 116.1%) (see Suppl. Fig. 1, available online at www.thrombosis-online.com). The anti-FXa activity of these subjects did not significantly change in at least two independent determinations with different samples. We sequenced 15,375 bp of the SERPINC1 gene, including the potential promoter, exons, introns


and the 3’ region (Fig. 1A). Comparison with the reference sequence revealed 18 polymorphisms and a new g.2085 T>C genetic change (Fig. 1 and Suppl. Table 3, available online at www.thrombosis-online.com). The last modification, identified in a subject with high antithrombin levels (H1), located in the potential promoter region, 228 bp upstream from the translation initiation codon (Fig. 1A). Interestingly, the g.2085 T>C transition would have potential transcriptional consequences, as it eliminates GATA-1 and c-Myb binding sites and creates a new site for AP-1 (Fig. 1B). The genotyping of this modification in 131 healthy subjects by HRM/Razor probe method (see Suppl. Fig. 2, available online at www.thrombosis-online.com) confirmed that this genetic change is a polymorphism. Actually, during this study, this change was already described as a SNP in the HapMap (rs78972925). We identified the g.2085 T>C transition in heterozygous state in seven out of 131 subjects (5%). However, this polymorphism only showed a trend to be associated with high anti-FXa activity (99 ± 11% in carriers vs. 96 ± 9% in non-carriers;p= 0.375) (Table 1). We also studied the potential functional effect of the other polymorphisms identified in this study by correlating genotype with anti-FXa activity in healthy subjects. This study was restricted to those polymorphisms identified only in individuals with high or

© Schattauer 2012


–170 bp from initiation codon respectively. Numbering is based on genomic DNA sequence, nucleotide +1 corresponds to the A of the ATG initiation codon of the Homo sapiens antithrombin gene (GenBank accession number NG_012462.1).

sation in the SERPINC1 locus of genetic variations identified in this study. The subjects carrying these variations are also indicated. S: small nucleotide series; L: large nucleotide series. B) Potential transcriptional binding sites affected by the g.2085 T>C and g.2143 C>G modifications located –228 and

Figure 1: Genetic variations identified in the SERPINC1 gene. A) Locali-

5

Figure 2: Identification and functional characterisation of the g.2143 C>G modification. A) Electropherogram of the mutation identified in Family 2. B) Family 2 pedigree. Proband is pointed by an arrow. Carriers of the g.2143 C>G modification are represented with the semi filled symbol. ND:

not determined. Plasma anti-FXa activities (%) are also shown. C) Luciferase reporter assay of SERPINC1 promoter activity with the wild-type and mutated g.2143G allele in HepG2 and PLC-PRF-5 cell lines. Results were obtained from five independent experiments.

low antithrombin levels. As shown in Table 1, only rs2227589, located in intron 1, was significantly associated with antithrombin levels. Carriers of the polymorphic allele had a moderately reduced anti-FXa activity in plasma (16). Haplotype analysis revealed 7 frequent haplotypes, although none of them seems to be significantly associated with anti-FXa activity (Suppl. Figs. 3 and 4, available online at www.thrombosis-online.com). Finally, our sequence analysis revealed three new alleles of the rs57195053. This polymorphism was originally described as a deletion of an ATT triplet from the sequence of 14 tandem triplets on intron 5. We detected alleles with 5, 11 and 16 triplets. This result together with the description of 11 new polymorphisms affecting this location (rs72125649, rs71708020, rs71799889, rs71117551, rs71563245, rs66670532, rs72352633, rs10598596, rs72304917, rs67446132 and rs57195053), supports a single polymorphism with high allelic variability (at least from 5 to 16 ATT repeats).

two patients (P1 and P7) carried two linked polymorphisms (rs2227588 and rs61827938) located in the promoter region. However, none of these polymorphisms modified potential binding sites for transcription factors. Moreover, the functional effect of these polymorphisms was discarded after genotyping in 126 blood donors (Table 1).

Families with antithrombin deficiency carrying neither gross gene deletion nor mutation affecting exons or flanking regions

Family 1 had severe family history of thrombosis and a type I antithrombin deficiency. Three sisters with deep venous thrombosis and recurrent events showed 50–56% of functional antithrombin activity and similar antigen levels. Multiplex Ligation-dependent Probe Amplification (MLPA) analysis and sequencing of both strands of exons and flanking regions were done in two affected members, but no significant genetic change was discovered. Moreover, we did not identify mutations in the rest of the SERPINC1 gene, including 1,500 bp from the promoter region and 1,000 bp from 3’ region. In contrast, three members of Family 2 had a moderate type I antithrombin deficiency, although no one has developed thrombotic events (Fig. 2A). The single relevant genetic change identified in the proband of Family 2 was a point mutation g.2143 C>G, located 170 bp upstream from the translation initiation codon. In silico prediction suggested that this change could eliminate a potential binding site for the LBP-1 regulatory factor (Fig. 1).

Thrombosis and Haemostasis 107.3/2012 © Schattauer 2012


de la Morena-Barrio et al. First mutation in the promoter gene of antithrombin

The clinical features and anti-FXa activity of these patients are shown in Suppl. Table 4 (available online at www.thrombosis-online.com). In these patients we only sequenced the coding region, flanking introns and the potential promoter region. We only detected three polymorphisms (Suppl. Table 4, available online at www.thrombosis-online.com). In addition to the rs3138521,

Genetic analysis in thrombophilic patients with moderate antithrombin deficiency


Then, we genotyped this transversion in 103 healthy blood donors by PCR-allelic specific restriction assay with Bsa JI, but no one carried it, supporting that this change is not a common polymorphism. In contrast, the mutation was present in the three membersof Family 2 with low antithrombin levels ( Fig. 2A). In order to further sustain the functional relevance of the g.2143 C>G transversion, luciferase reporter assays were performed. The plasmid containing the mutated promoter was associated with 30–50% reduction of luciferase activity in two different cell lines (HepG2 and PLC-PRF-5) ( Fig. 2B).

Discussion

In the effort to find thrombotic risk factors, the identification of new elements able to affect the levels of key haemostatic molecules, such as antithrombin, is motivating. In addition to rare mutations causing severe antithrombin deficiency that significantly increase the risk of thrombosis (17), moderate reduction in antithrombin levels has recently been suggested to mildly increase the thrombotic risk (personal communication with Dr. I. Martinelli). Therefore, it would be of interest to identify new genetic variations associated with low antithrombin levels. This search has rendered excellent results in patients with congenital antithrombin deficiency, as the causative mutation is identified in a high percentage of cases (up to 90% according to recent large studies) (18). However, all

Thrombosis and Haemostasis 107.3/2012 © Schattauer 2012


of g.2143 C>G in potential promoter regulatory regions of the SERPINC1 gene identified by DNase protection assay. Sequences highlighted in grey (A-C) were described by Fern- andez-Rachubinski et al. (10) while regions underlined (I-IV) were described by Tremp et al. (10, 26). The interacting transcription factors identified are indicated below each sequence.

Figure 3: Localisation

genetic defects identified in these studies are gross gene deletionsor mutations affecting exons or flanking regions. The absence of genetic modifications associated with reduced levels of this key anticoagulant through regulatory mechanisms is surprising, probably reflecting that mutations affecting these regions are very rare or have small functional consequences. Alternatively, it can also be explained by the few studies that have tried to find them, probably because there are few regulatory regions identified in the SERPINC1 gene. So far, only one report from our group has reported a very mild effect associated to a common polymorphism in intron 1 (16), which also explains its negligible thrombotic consequences (19, 20). This study aimed to identify new genetic modifications associated with antithrombin levels through a regulatory mechanism, by sequencing the SERPINC1 gene in three groups of samples with different antithrombin levels. Moreover, the identification of a functional change will also help to discover or authenticate regions with regulatory relevance in this gene. The first approach was to dissect the genetic architecture of the SERPINC1 locus in healthy subjects with extreme antithrombin levels within the normal range (75% and 115%). The anti-FXa activity, the method widely used in thrombophilic test to detect antithrombin deficiency, shows a great variability with normal distribution (70–130%) (21). Factors such as sex, body mass index, oral contraceptives or race have been associated with antithrombin variability (22), but genetic factors must also play a relevant role according to the high heritability of this trait (h= 0.486) (23). We found a polymorphism that deserved our attention. The g.2085

de la Morena-Barrio et al. First mutation in the promoter gene of antithrombin 7

What is known about this topic?

- Antithrombin is a key anticoagulant with wide variability in normal population. The high heritability of this trait sustains a role for genetic factors. - Mutations in exons and flanking regions of SERPINC1, the gene encoding antithrombin, caused antithrombin deficiency, but only one polymorphism, rs2227589 located in intron 1, has mild functionalconsequences. - DNase footprinting assays showed four potential regulatory regions in the 200 bp upstream the start codon.

What does this paper add?

- We identified the first mutation in the SERPINC1 promoter associated with antithrombin deficiency. This result recommends the analysis of this region in the molecular characterisation of subjects with antithrombin deficiency, particularly those with type I deficiency without mutation in the coding region of SERPINC1. - Our study provides new evidences of the reduced genetic variability in SERPINC1 and the minor functional consequences of common polymorphisms. - Our data support the existence of modulating genes that might beinvolved in antithrombin variability.

levels identified in carriers, but significantly impaired it. This functional effect was further confirmed by reporter assays in different cell lines. Moreover, our result verify the regulatory relevance of the region of at least 200 bp upstream the ATG start codon. The g.2143 C>G modification is located in one of the four protected regions identified in two independent reports (10, 26). Thus, it is located at the end of region C proposed by Fernandez-Rachubinski et al.(10), and in the middle of region II according to the nomenclature of Tremp et al. (26) (Fig. 3). Interestingly, it has been suggested that the hepatic nuclear factor 4 (HNF4) interacts with this region (10) (Fig. 3). Therefore, our results support that this region, and probably the whole promoter region, should be included in the molecular analysis of subjects with antithrombin deficiency, particularly in those subjects with type I deficiency and without mutation in the coding region of SERPINC1.

Acknowledgements

We acknowledge L. Velázquez and N. Bohdan for their excellent technical assistance.

Conflict to interest

None declared.

T>C change identified in a subject with high antithrombin levels could potentially modify the binding of different transcription factors (Fig. 1B). However, our study revealed that this polymorphism had no significant functional effect. The low allelic frequency of this polymorphism (0.027) encourages confirming these results in further studies including more subjects. Furthermore, despite its small size, our study provides evidence of a reduced genetic variability in SERPINC1 and the minor functional consequences of common polymorphisms, particularly those affecting coding regions. This can be explained by the extraordinary functional and conformational sensitivity of this molecule to even minor modifications that would result in a classical antithrombin deficiency (24). These data, together with the high heritability of antithrombin levels (23) support the existence of modulating genes that might be involved in antithrombin variability, and could play a role in the course or severity of thrombotic disorders (25). These genes might encode for transcriptional factors, molecules involved in folding or secretion, enzymes implicated in post-translational modifications, proteases, or miRNAs. An effort to identify these potential modulating genes of antithrombin has to be done. However, we cannot exclude the presence of regulatory elements located up or down stream the region studied. Actually, the identification of a family (Family 1), with type I antithrombin deficiency and severe thrombosis who had neither gross deletion nor mutation in the 15,925 bp of the SERPINC1 gene strongly suggests the existence of remote regulatory regions. Finally, we identified the first mutation in the promoter region associated with antithrombin deficiency. Certainly, the g.2143 C>G change identified in Family 2 does not abolish the transcription of the mutated allele, as demonstrated by the antithrombin

© Schattauer 2012

References

1. Abildgaard U. Antithrombin--early prophecies and present challenges. Thromb Haemost 2007; 98: 97–104. 2. Rau JC, Beaulieu LM, Huntington JA, et al. Serpins in thrombosis, hemostasis and fibrinolysis. J Thromb Haemost 2007; 5 (Suppl 1): 102–115. 3. Bayston TA, Lane DA. Antithrombin: molecular basis of deficiency. Thromb Hae- most 1997; 78: 339–343. 4. Garcia de Frutos P. Mechanisms of thrombophilia. Thromb Haemost 2007; 98: 485–487. 5. Lane DA, Bayston T, Olds RJ, et al. Antithrombin mutation database: 2nd (1997) update. For the Plasma Coagulation Inhibitors Subcommittee of the Scientific and Standardization Committee of the International Society on Thrombosis and Haemostasis. Thromb Haemost 1997; 77: 197–211. 6. Olds RJ, Lane DA, Chowdhury V, et al. Complete nucleotide sequence of the antithrombin gene: evidence for homologous recombination causing thrombophilia. Biochemistry 1993; 32: 4216–4224. 7. Hernandez-Espinosa D, Ordonez A, Vicente V, et al. Factors with conformational effects on haemostatic serpins: implications in thrombosis. Thromb Haemost 2007; 98: 557–563. 8. Bock SC, Levitan DJ. Characterization of an unusual DNA length polymorphism 5' to the human antithrombin III gene. Nucleic Acids Res 1983; 11: 8569–8582. 9. Prochownik EV. Relationship between an enhancer element in the human antithrombin III gene and an immunoglobulin light-chain gene enhancer. Nature 1985; 316: 845–848. 10. Fernandez-Rachubinski FA, Weiner JH, Blajchman MA. Regions flanking exon 1 regulate constitutive expression of the human antithrombin gene. J Biol Chem 1996; 271: 29502–29512. 11. Ochoa A, Brunel F, Mendelzon D, et al. Different liver nuclear proteins binds to similar DNA sequences in the 5' flanking regions of three hepatic genes. Nucleic Acids Res 1989; 17: 119–133. 12. Sheffield WP, Wu YI, Blajchman MA. Antithrombin: Structure and Function. In: Molecular Basis of Thrombosis and Hemostasis.New York: Marcel Dekker, Inc.; 1995. pp. 355–377. 13. Corral J, Huntington JA, Gonzalez-Conejero R, et al. Mutations in the shutter region of antithrombin result in formation of disulfide-linked dimers and severe venous thrombosis. J Thromb Haemost 2004; 2: 931–939. 14. Corral J, Aznar J, Gonzalez-Conejero R, et al. Homozygous deficiency of heparin cofactor II: relevance of P17 glutamate residue in serpins, relationship with con-



8

formational diseases, and role in thrombosis. Circulation 2004; 110: 1303–1307. 15. Sole X, Guino E, Valls J, et al. SNPStats: a web tool for the analysis of association studies. Bioinformatics 2006; 22: 1928–1929. 16. Anton AI, Teruel R, Corral J, et al. Functional consequences of the prothrombotic SERPINC1 rs2227589 polymorphism on antithrombin levels. Haematologica 2009; 94: 589–592. 17. Maclean PS, Tait RC. Hereditary and acquired antithrombin deficiency: epidemiology, pathogenesis and treatment options. Drugs 2007; 67: 1429–1440. 18. Luxembourg B, Delev D, Geisen C, et al. Molecular basis of antithrombin deficiency. Thromb Haemost 2011; 105: 635–646. 19. Bezemer ID, Bare LA, Doggen CJ, et al. Gene variants associated with deep vein thrombosis. J Am Med Assoc 2008; 299: 1306–1314. 20. Austin H, De SC, Lally C, et al. New gene variants associated with venous thrombosis: a replication study in White and Black Americans. J Thromb Haemost 2011; 9: 489–495.

21. Tait RC, Walker ID, Islam SI, et al. Influence of demographic factors on antithrombin III activity in a healthy population. Br J Haematol 1993; 84: 476–480. 22. Conlan MG, Folsom AR, Finch A, et al. Antithrombin III: associations with age, race, sex and cardiovascular disease risk factors. The Atherosclerosis Risk in Com- munities (ARIC) Study Investigators. Thromb Haemost 1994; 72: 551–556. 23. Souto JC, Almasy L, Borrell M, et al. Genetic determinants of hemostasis phenotypes in Spanish families. Circulation 2000; 101: 1546–1551. 24. Corral J, Vicente V, Carrell RW. Thrombosis as a conformational disease. Haema- tologica 2005; 90: 238–246. 25. Westrick RJ, Ginsburg D. Modifier genes for disorders of thrombosis and hemostasis. J Thromb Haemost 2009; 7 (Suppl 1): 132–135. 26. Tremp GL, Duchange N, Branellec D, et al. A 700-bp fragment of the human antithrombin III promoter is sufficient to confer high, tissue-specific expression on human apolipoprotein A-II in transgenic mice. Gene 1995; 156: 199–205.

Thrombosis and Haemostasis 107.3/2012 © Schattauer 2012 Downloaded from www.thrombosis-online.com on 2012-02-14 | ID: 1000465936 | IP: 134.58.253.55

de la Morena-Barrio et al. First mutation in the promoter gene of antithrombin

Letters to the Editor 2625 Congenital disorder of glycosylation (PMM2-CDG) in a patient with antithrombin deficiency and severe thrombophilia M . E . D E L A M O R E N A - B A R R I O , * T . S . S E V I V A S , < I . M A R T I N E Z - M A R T I N E Z , * A . M I N

˜ A N O , * V . V I C E N T E , * J . J A E K E N ₃ and J . C O R R A L *

*Centro Regional de Hemodonacio´n, University of Murcia, Regional Campus of International Excellence ₃Campus Mare Nostrum₃, Murcia, Spain; <Department of Hematology, Centro Hospitalar de Coimbra, Portugal; and ₃Center for Metabolic Diseases, Universitair Ziekenhuis Gasthuisberg, Leuven, Belgium To cite this article: de la Morena-Barrio ME, Sevivas TS, Martínez-Martínez I, Miñano A, Vicente V, Jaeken J, Corral J. Congenital disorder of glycosylation (PMM2-CDG) in a patient with antithrombin deficiency and severe thrombophilia. J Thromb Haemost 2012; 10: 2625–7.

Antithrombin deficiency was the first congenital thrombophilic factor identified [1]. Since then, protein C and protein S deficiencies, as well as specific mutations in the factor V and prothrombin genes, have joined antithrombin deficiency as genetic defects predisposing patients to venous thrombosis. However, it is unlikely that new common polymorphisms could play a key role in venous thrombosis [2]. Therefore, rare mutations affecting different elements of the hemostatic system are emerging candidates [3,4]. The analysis of antithrombin in a thrombophilic patient with antithrombin deficiency has allowed us to identify a particular thrombophilic defect that, affecting the N-glycosylation pathway, indirectly disturbs this key anticoagulant serpin.

We report on an 18-year-old male with recurrent left-leg venous thrombosis. The first event happened at 11 years of age. Oral anticoagulation with warfarin was established for 5 months. A second event occurred at the same location at 12 years. Oral anticoagulation with warfarin was re-estab-lished. Surprisingly, a new thrombotic event took place 6 months later, under stable oral anticoagulant therapy with an International Normalized Ratio (INR) of 2–3. A fourth thrombotic event was diagnosed at 17 years, again under oral anticoagulant therapy. Currently, the patient is being treated with higher warfarin doses (INR 2.5–3.5).

No familial history of thrombosis was reported. Thromb-ophilic analysis revealed no prothrombotic polymorphisms (FV Leiden or prothrombin G20210A). No antiphospholipid antibodies were detected. Protein C and protein S levels were within the normal range. The only thrombophilic defect identified was an antithrombin deficiency (anti-FXa activity of 67%). Antigen levels of antithrombin were also low (60%), suggesting a spontaneous heterozygous type I antithrombin Correspondence: Dr Javier Corral, University of Murcia, Centro Regional de Hemodonacio´n, Ronda de Garay S/N, Murcia 30003, Spain. Tel.: +34 968341990; fax: +34 968261914. E-mail: [email protected] DOI: 10.1111/jth.12031. Received 4 September 2012, accepted 15 October 2012 ₃ 2012 International Society on Thrombosis and Haemostasis

deficiency. However, the sequencing of the promoter, as well as the coding and flanking regions of the SERPINC1 gene, [5] revealed no mutation. Large gene deletions were excluded by multiplex ligation-dependent probe amplification (MLPA) studies. Interestingly, electrophoretic analysis of plasma anti-thrombin performed under either denaturing or native con-ditions [6] revealed the presence of an abnormal antithrombin (Fig. 1A). The increased mobility in SDS suggested abnormal post-translational modifications. We evaluated impaired N-glycosylation, the only post-translational modification de-scribed in antithrombin that also affects the function and clearance of this anticoagulant serpin [7]. N-glycan removal with N-glycosidase F [6] equalized the mobility of both the major and the minor antithrombin species, suggesting an N-glycosylation defect for ₃ 20% of the antithrombin molecules in the plasma of the proband (Fig. 1B). This abnormal antithrombin was not the b-glycoform as, among other findings, this defect was not specific for antithrombin. Thus, smaller forms of a1-antitrypsin were also detected in the plasma of this patient (Fig. 1C). These biochemical and clinical data suggested a congenital disorder of glycosylation (CDG), a category that comprises a broad group of autoso-mal recessive disorders [8]. Actually, the antithrombin and a1-antitrypsin electrophoretic profiles were very similar to those shown by a patient with the commonest CDG, PMM2-CDG, with mutations in PMM2, the gene encoding phosphoman-nomutase 2. This is a cytosolic enzyme that catalyzes the conversion of mannose 6-phosphate to mannose 1-phosphate, a substrate for GDP-mannose synthesis, required in N-glycosylation [8] (Fig. 1C). HPLC analysis of transferrin glycoforms pointed to a CDG-I diagnosis by the presence of asialotransferrin and a significant increase in the amount of disialotransferrin (Fig. 1D). We identified two mutations in PMM2: one of maternal origin in exon 5, and the other in exon 8, responsible for the p.R141H and p.C241S substitu-tions, respectively. Unfortunately, we do not have samples from the father. The same combination has been reported previously in a PMM2-CDG patient with a mild clinical phenotype [9]. In addition, the proband had mental retarda-tion without ataxia, strabismus, a thyroid nodule, and kyphosis. The digestive, hepatic, renal and cardiovascular systems were normal.

A Native SDS-red

58 kDa 55 kDa

B SDS-red

58 kDa 55 kDa

Proband def AT Control

Proband Control Proband Control

49 kDa

D 1.2

1.0

(mAU

) 0.8

0.6

Abs

orba

nce

0.4 Mother

0.2

0.0

–0.2

0 5 10 15 20 25

Time (min)

1.2

1.0

0.8

N-glycosidase F – +

(mAU

) 0.6

C

SDS-red

α1-AT

52 kDa 50 kDa 48 kDa

AT 58 kDa 55 kDa

PMM2-CDG

Proband Mother Control

Abso

rban

ce

0.4

0.2

0.0

–0.2

0 5 10 15 20 25

Time (min)

1.2

1.0

(mAU

) 0.8

0.6

Abs

orba

nce

0.4 PMM2-CDG

0.2

0.0

–0.2

0 5 10 15 20 25

Time (min)

Fig. 1 Identification of PMM2-CDG in the proband. (A) Native and SDS-PAGE showing the normal (arrow) and abnormal (dashed arrow) anti-thrombin present in plasma of the proband, a patient with congenital type I antithrombin deficiency (AT def) (p.R149X) and a control. (B) SDS-PAGE and Western blot of ntithrombin from plasma untreated (–) or treated (+) with N-glycosidase F. (C) SDS-PAGE pattern of a1-antitrypsin (a1-AT) and antithrombin (AT) in the proband, his mother, a control, and a PMM2-CDG patient (p.F119L and p.R141H). Arrows mark the abnormal glycoforms of these molecules. (D) HPLC profile of transferrin glycoforms of the proband, his mother, and a PMM2-CDG patient (p.F119L and p.R141H). Up to 55% of PMM2-CDG patients show acute events such as ₃stroke-like₃ episodes and ischemic cerebral vascular acci-dents, and at least 10 cases with either venous or arterial thrombosis have been described [10]. Acute vascular events occurred in patients younger than 15 years, especially during fever and prolonged immobilization [10]. These accidents can

be recurrent in the same patient [10]. Only a few studies have evaluated hemostatic factors [10,11] and platelets [12] in PMM2-CDG with the aim of identifying the mechanism(s) involved in the development of these vascular events. PMM2-CDG patients consistently have antithrombin deficiency [10,11], with abnormal hypoglycosylated antithrombin mole-cules [10]. This might contribute to the increased risk of thrombotic events described in these patients. However, there

Proband

2625 Letter to the editor

are no differences in antithrombin levels between patients with and without vascular events [10].

This is the first report on a PMM2-CDG patient with documented recurrent venous thrombosis. Moreover, the patient suffered from thrombotic events without a clear triggering factor, such as fever, catheterization, or immobiliza-tion. Finally, two recurrent thrombotic events occurred under oral anticoagulant therapy. This unusual clinical severity suggests that hemostatic or vascular defects other than those described in PMM2-CDG patients (or additional unknown thrombophilic anomalies) might be present in this patient. This deserves further study.

This case report also highlights another important conclu-sion concerning the diagnostic procedure in patients with antithrombin deficiency. Direct sequencing and MLPA anal-ysis of SERPINC1 in patients with thrombosis and antithrom-bin deficiency is highly effective. A recent study including 234 cases with antithrombin deficiency found a high mutation detection rate for SERPINC1 (83.5%) [13]. Our study suggests that a diagnosis of PMM2-CDG might be suspected among cases with antithrombin deficiency and no mutations in SERPINC1, particularly in those cases without a family history of thrombosis, as well as in cases with a lower molecular weight plasma antithrombin. Disclosure of conflict of interests This study was supported by 04515/GERM/06, PI12/00657 (ISCIII and FEDER), and RECAVA RD06/0014/0039 (ISCIII and FEDER). M. E. de la Morena-Barrio is a holder of a predoctoral research grant from ISCIII (FI09/00190). I. Martínez-Martínez is a researcher from Fundación para la Formación e Investigación Sanitarias de la Región de Murcia.

Acknowledgements We are indebted to the patient and his family for their collaboration in this study. References 1 Egeberg O. Inherited antithrombin deficiency causing thrombophilia.

Thromb Diath Haemorrh 1965; 13: 516–30.

Letters to the Editor 2627 2 Heit JA, Armasu SM, Asmann YW, Cunningham JM, Matsumoto

ME, Petterson TM, de Andrade M. A genome-wide association study of venous thromboembolism identifies risk variants in chromo-

somes 1q24.2 and 9q. J Thromb Haemost 2012; 10: 1521–31. 3 Miyawaki Y, Suzuki A, Fujita J, Maki A, Okuyama E, Murata M,

Takagi A, Murate T, Kunishima S, Sakai M, Okamoto K, Matsushita

T, Naoe T, Saito H, Kojima T. Thrombosis from a prothrombin mutation conveying antithrombin resistance. N Engl J Med 2012; 366: 2390–6.

4 Simioni P, Tormene D, Tognin G, Gavasso S, Bulato C, Iacobelli NP, Finn JD, Spiezia L, Radu C, Arruda VR. X-linked thrombophilia with a mutant factor IX (factor IX Padua). N Engl J Med 2009; 361: 1671– 5.

5 de la Morena-Barrio ME, Anton AI, Martinez-Martinez I, Padilla J, Minano A, Navarro-Fernandez J, Aguila S, Lopez MF, Fontcuberta J, Vicente V, Corral J. Regulatory regions of SERPINC1 gene: iden-

tification of the first mutation associated with antithrombin deficiency. Thromb Haemost 2012; 107: 430–7.

6 Martinez-Martinez I, Johnson DJ, Yamasaki M, Navarro-Fernandez J, Ordonez A, Vicente V, Huntington JA, Corral J. Type II anti-thrombin deficiency caused by a large in-frame insertion. Structural, functional and pathological relevance. J Thromb Haemost 2012; 10: 1859–66.

7 McCoy AJ, Pei XY, Skinner R, Abrahams JP, Carrell RW. Structure

of beta-antithrombin and the effect of glycosylation on antithrombin₃s

heparin affinity and activity. J Mol Biol 2003; 326: 823–33. 8 Jaeken J. Congenital disorders of glycosylation (CDG): it₃s (nearly)

all in it. J Inherit Metab Dis 2011; 34: 853–8.

9 Gru¨newald S, Schollen E, Van Schaftingen E, Jaeken J, Matthijs G.

High residual activity of PMM2 in patients₃ fibroblasts: possible pitfall in the diagnosis of CDG-Ia (phosphomannomutase deficiency). Am J Hum Genet 2001; 68: 347–54.

11 Arnoux JB, Boddaert N, Valayannopoulos V, Romano S, Bahi-

Buisson N, Desguerre I, de Keyzer Y, Munnich A, Brunelle F, Seta N, Dautzenberg MD, de Lonlay P. Risk assessment of acute vascular events in congenital disorder of glycosylation type Ia. Mol Genet Metab 2008; 93: 444–9.

12 Young G, Driscoll MC. Coagulation abnormalities in the carbohy-

drate-deficient glycoprotein syndrome: case report and review of the literature. Am J Hematol 1999; 60: 66–9.

13 Van Geet C, Jaeken J, Freson K, Lenaerts T, Arnout J, Vermylen J,

Hoylaerts MF. Congenital disorders of glycosylation type Ia and IIa are associated with different primary haemostatic complications. J Inherit Metab Dis 2001; 24: 477–92.

14 Caspers M, Pavlova A, Driesen J, Harbrecht U, Klamroth R, Kadar

J, Fischer R, Kemkes-Matthes B, Oldenburg J. Deficiencies of antithrombin, protein C and protein S – practical experience in genetic analysis of a large patient cohort. Thromb Haemost 2012; 108: 247–57.

₃ 2012 International Society on Thrombosis and Haemostasis

Appendix II

149

Diffusion of other scientific results

Appendix II

Appendix II

150

Appendix II

151

This appendix includes other results that have been published in peer-reviewed

journals.

1. López-Contreras AJ, Sánchez-Laorden BL, Ramos-Molina B, de la Morena

ME, Cremades A, Peñafiel R. Subcellular localization of antizyme inhibitor 2 (AZIN2)

in mammailian cells: Influence of intrinsic sequences and interaction with antizymes. J.

Cell Biochem. 107; 732-740; 2009. Impact Factor: 3.122.

2. Martínez-Martínez I, Ordóñez A, Pedersen S, de la Morena-Barrio ME,

Navarro-Fernández J, Kristensen SR, Miñano A, Padilla J, Vicente V, Corral J. Heparin

affinity of factor VIIa: Implications on the physiological inhibition by antithrombin and

clearance of recombinant factor VIIa. Thrombosis Research. 127 (2); 154-160; 2011.

Impact Factor: 2.372.

3. Guerrero JA, Teruel R, Martínez C, Arcas I, Martínez-Martínez I, de la

Morena-Barrio ME, Vicente V, Corral J. Protective role of antithrombin in mouse

models of liver injury. J Hepatol. 2012 Nov; 57(5):980-6. Impact factor: 9.264.

4. Martínez-Martínez I, Navarro-Fernández J, Østergaard A, Gutiérrez-Gallego R,

Padilla J, Bohdan N, Miñano A, Pascual C, Martínez C, de la Morena-Barrio ME,

Aguila S, Pedersen S, Kristensen SR, Vicente V, Corral J. Amelioration of the severity

of heparin-binding antithrombin mutations by posttranslational mosaicism. Blood. 2012

Jul 26;120(4):900-4. Impact factor: 9.898.

5. Martínez-Martínez I, Navarro-Fernández J, Aguila S, Miñano A, Bohdan N, de

La Morena-Barrio ME, Ordóñez A, Martínez C, Vicente V, Corral J. The infective

polymerization of conformationally unstable antithrombin mutants may play a role in

the clinical severity of antithrombin deficiency. Mol Med. 2012 Jul 18;18(1):762-70.

Impact factor: 3.76.

6. Águila S, Martínez-Martínez I, Collado M, Llamas P, Antón AI, Martínez-

Redondo C, Padilla J, Miñano A, de la Morena-Barrio ME, García-Avello A, Vicente

V, Corral J. Compound heterozygosity involving Antithrombin Cambridge II

(p.Arg416Ser) in antithrombin deficiency. Thromb. Haemost. 2013 Jan 17;109(3).

[Epub ahead of print]. Impact Factor: 5.044.

Appendix II

152

Documents

Un sistema circulatorio cerrado y de alta presión precisa de un